UHI Thesis - pdf download summary
Quantitative analysis of active compounds found in Arnica montana and Arnica chamissonis in relation to varied provenance, environmental and agronomic factors.
Barron-Majerik, Elizabeth
DOCTOR OF PHILOSOPHY (AWARDED BY OU/ABERDEEN)
Award date: 2011
Awarding institution: The University of Edinburgh
Link URL to thesis in UHI Research Database
General rights and useage policy Copyright,IP and moral rights for the publications made accessible in the UHI Research Database are retained by the author, users must recognise and abide by the legal requirements associated with these rights. This copy has been supplied on the understanding that it is copyright material and that no quotation from the thesis may be published without proper acknowledgement, or without prior permission from the author.
Users may download and print one copy of any thesis from the UHI Research Database for the not-for-profit purpose of private study or research on the condition that:
1) The full text is not changed in any way 2) If citing, a bibliographic link is made to the metadata record on the the UHI Research Database 3) You may not further distribute the material or use it for any profit-making activity or commercial gain 4) You may freely distribute the URL identifying the publication in the UHI Research Database Take down policy If you believe that any data within this document represents a breach of copyright, confidence or data protection please contact us at [email protected] providing details; we will remove access to the work immediately and investigate your claim.
Download date: 08. Oct. 2021 Quantitative analysis of active compounds found in Arnica
montana and Arnica chamissonis in relation to varied
provenance, environmental and agronomic factors.
A thesis presented for the degree of Doctor of Philosophy at the University of
Aberdeen
June 2010
Elizabeth J. Barron
BSc (Hons) Plant Science, University of Edinburgh
Arnica montana L.
Inverness College UHI Longman Campus Inverness College Highlands IV1 1SA
Declaration
I, Elizabeth Jane Barron, confirm that I composed the thesis, that it has not been accepted in any previous application for a degree, that the work is my own, and that all quotations have been distinguished by quotation marks and the sources of information specifically acknowledged.
Signature:
Date:
i
Acknowledgements
Firstly, thanks go to my family. Without their unconditional support, love and wine, none of this would have been possible. Thank you also to my friends who have supported me throughout this project. Particularly, Katie, Gordon, Wendy, Matthew and Chantel: I owe you. Jan, you were right about the film, but thanks for sticking by me nonetheless.
This was a part-time PhD, which meant it took a lot longer than most projects to be completed. Credit is due to my very patient supervisors, Melanie and Peter (and
Paddy who retired before I could finish), who miraculously maintained enthusiasm for this project, even when mine was waning. Thank you to the Agronomy Team at
Orkney, particularly John Wishart whose help was indispensable to this project.
Thank you to Kenny Boyd and Sandy Gray for your advice and guidance and thanks also to my very understanding line manager Janet Mahon, who expertly found the perfect balance between support, encouragement and blatant threats.
But above all this is for my dad, who isn’t here to see me finish, but it is thanks to him that I did.
ii
Abstract
The northern hemisphere plant Arnica montana L. can be found across Europe on high alpine slopes with acidic soils. The flowers are well documented as a source of natural products, particularly, active compounds associated with anti- inflammatory properties. The high demand for this extract has meant that wild plant numbers have reduced dramatically and now in many countries, it is illegal to harvest A. montana from the wild. As a result, flowers from the non-endangered A. chamissonis Less. have been employed for the production of extracts with similar properties.
A. montana and A. chamissonis were both grown during the period 2004 to 2007 in the North of Scotland (Orkney) in order to determine whether this environment was condusive to high yield. The qualitative and quantitative effects of weeding, fertiliser, mulch, seed source and species on yield were also studied. This research found that A. montana had high quality extract but was prone to crown rot and was less robust than initial trials suggested. Conversely, while A. chamissonis grew vigorously, the extract contained approximately a third of the sesquiterpene lactone content.
iii Contents iv
Figures xi
Tables xxi
List of Abbreviations xxv
1. Introduction 1
1.1 Research context 1
1.2 Arnica extract 2
1.3 The market 7
1.4 The taxonomy of Arnica 9
1.4.1 Arnica montana 12
1.4.1.1 Distribution and growth 12
1.4.1.2 Description 14
1.4.2 Arnica chamissonis 16
1.4.2.1 Distribution and growth 16
1.4.2.2 Description 17
1.5 Pests and grazing 19
1.6 Conservation and cultivation 21
1.7 Alternative agricultural crops and the Highlands and Islands 24
1.8 The chemistry of Arnica 29
1.8.1 Terpenes 31
1.8.2 Sesquiterpenes 33
1.8.3 Sesquiterpene lactones 34
1.8.3.1 Classification of sesquiterpene lactones 35
1.8.3.2 Sesquiterpene lactones and chemosystematics 39
iv 1.8.3.3 Sesquiterpene lactones in Arnica 42
1.8.3.4 The role of sesquiterpene lactones in plants 48
1.8.3.5 The medical uses of sesquiterpene lactones 49
1.8.3.6 Chemical synthesis of sesquiterpene lactones 55
1.8.4 Other constituents of Arnica extract 55
1.8.5 Toxicology 57
1.9 Discussion 59
1.10 Research question 61
1.11 Aim 61
1.12 Objectives for this study 61
2. The extraction, analysis and identification of sesquiterpene lactones 63
2.1 Introduction 63
2.1.1 Flower material 63
2.1.2 Extraction 64
2.1.3 Chemical analysis 73
2.1.4 Compound identification 74
2.1.5 Quantification 80
2.2 Methodology 83
2.2.1 Flower material 83
2.2.2 Extraction methods 84
2.2.2.1 Hydrodistillation 84
2.2.2.2 Solvent extraction 86
2.2.3 GC-MS methodology 87
2.2.4 Compound identification 88
2.2.5 Quantification of sesquiterpene lactones 89
v 2.3 Results 92
2.3.1 Calibration curve 92
2.3.2 Hydrodistillation extraction results 94
2.3.3 Results of essential oil analysis 94
2.3.4 Solvent extraction compound identification of A. montana 102
2.3.4.1 Peak 1 103
2.3.4.2 Peak 2 104
2.3.4.3 Peak 3 106
2.3.4.4 Peak 4 107
2.3.4.5 Peak 5 109
2.3.4.6 Peak 6 110
2.3.4.7 Peak 7 111
2.3.4.8 Peak 8 112
2.3.4.9 Peak 9 113
2.3.4.10 Peak 10 115
2.3.4.11 Peak 11 116
2.3.4.12 Peak 12 117
2.3.4.13 Peak 13 118
2.3.4.14 Peak 14 119
2.3.5 Quantitative analysis of A. montana solvent extract 120
2.3.5.1 Commercial and Orcadian extract 120
2.3.5.2 Extract from different floral stages 124
2.3.6 Solvent extraction compound identification of A. chamissonis 128
2.3.6.1 Peak 15 128
vi 2.3.6.2 Peak 16 131
2.3.6.3 Peak 17 133
2.3.7 Quantitative analysis of A. chamissonis solvent extract 135
2.3.7.1 A. chamissonis floral extract 135
2.3.7.2 Extract from different floral stages 136
2.4 Discussion 138
3. Yield of A. montana and A. chamissonis over 6 years 145
3.1 Introduction 145
3.1.1 Yield: a balance between quality and quantity 145
3.1.2 The established trials 148
3.1.3 Harvesting method 152
3.1.4 The climate 156
3.1.5 The soil 160
3.2 Materials and Methods 162
3.2.1 The site 162
3.2.2 Meteorological data 163
3.2.3 Flower harvesting 164
3.2.4 Extraction 165
3.2.5 Analysis 165
3.2.6 Statistical analysis 165
3.3 Results 166
3.3.1 Soil 166
3.3.2 Climate 170
3.3.3 Yield 175
vii 3.3.4 Competition and herbivory 179
3.3.5 Correlations between yield and environmental conditions 184
3.3.6 Quantitative and qualitative analysis of floral extract 20 5
3.4 Discussion of the yield and floral extract results 219
4. The influence of fertiliser and weeding on yield 227
4.1 Introduction 227
4.2 Materials and Methods 236
4.2.1 The sites 236
4.2.1.1 AM3 236
4.2.1.2.AM5 238
4.2.2 Treatment regime 239
4.2.3 Harvesting 242
4.2.4 Extractions 243
4.2.5 Analysis 243
4.3 Results 245
4.3.1 Plant survival 245
4.3.2 Yield 248
4.3.2.1 AM3 248
4.3.2.2 AM5 250
4.3.3 Essential oil 253
4.3.4 Sesquiterpene lactone production 257
4.4 Discussion 262
5. The influence of soil and planting regime on yield 268
5.1 Introduction 268
5.2 Materials and Methods 275
viii 5.2.1 The sites 275
5.2.1.1 Mulch (AM7) 276
5.2.1.2 Ridges and mulch (ACM1) 278
5.2.1.3 Ridges and seed source (ACM2) 281
5.2.2 Chemical analysis 282
5.2.3 Statistical analysis 282
5.3 Results 283
5.3.1 Plant survival 283
5.3.1.1 AM7 283
5.3.1.2 ACM1 289
5.3.1.3 ACM2 290
5.3.2 Flower production 291
5.3.2.1 AM7 291
5.3.2.2 ACM1 291
5.3.2.3 ACM2 294
5.3.3 Sesquiterpene lactone production 294
5.3.3.1 AM7 294
5.3.3.2 ACM1 – A. montana 295
5.3.3.3 ACM1 – A. chamissonis 299
5.3.3.4 ACM2 – A. montana 299
5.3.3.5 ACM2 – A. chamissonis 303
5.4 Discussion 305
6. Discussion and conclusion 312
6.1 Aims and objectives 312
6.2 Trial parameters 312
ix 6.3 Oil yield, profile and sesquiterpene lactone content 315
6.3.1 A.chamissonis 315
6.3.2 A.montana 316
6.4 The effect of agronomic and environmental conditions on
crop yield and quality 319
6.4.1 Environmental conditions 319
6.4.2 Agronomic regimes 324
6.5 The ‘ideal’ growing regime for Arnica grown in Orkney 330
6.5.1 A.chamissonis 330
6.5.2 A.montana 332
6.6 The role of sesquiterpene lactones in Arnica 336
6.6.1 A.chamissonis 336
6.6.2 A.montana 337
6.7 The potential of Orkney grown A.montana as a commercial
high value extract crop 342
6.8 Recommendations 347
6.9 Conclusions 348
7. References 349
x Figures
Figure 1.1 Arnica montana 3
Figure 1.2 Phylogenetic interrelationships of the orders of flowering plants 11
Figure 1.3 Occurrence of A.montana 13
Figure 1.4 A. montana flowers 14
Figure 1.5 Floral diagram of A. montana 15
Figure 1.6 Achenes and pappus of A. montana 16
Figure 1.7 US distribution of A.chamissonis ssp .foliosa 17
Figure 1.8 A. chamissonis flowers 18
Figure 1.9 Achenes and pappus of A. chamissonis 19
Figure 1.10 Scottish Land Cover Map, 2000 26
Figure 1.11 Map of the Highlands with the Orkney Islands highlighted 27
Figure 1.12 Isoprene unit 31
Figure 1.13 Regular and irregular arrangement of isoprene units 32
Figure 1.14 Organisation of terpene biosynthesis in plants. 34
Figure 1.15 The α-methylene-γ-lactone ring 36
Figure 1.16 The β methyl of the ambrosanolide type and α methyl group of the helanolides precursor 36
Figure 1.17 Hypothetical synthesis routes of different sesquiterpene lactones 38
Figure 1.18 Evolutional relations of tribes of Asteraceae 41
Figure 1.19 Diagram of helenalin, dihydrohelenalin and their esters 45
Figure 1.20 The Michael addition process 49
Figure 1.21 Tenulin 53
Figure 1.22 The product of reacting glutathione with the α-methylene
xi group of the γ-lactone and the double bond of the cyclopentenone ring 53
Figure 2.1 Sample sites of North American A.chamissonis 72
Figure 2.2 McLafferty rearrangements 76
Figure 2.3 Proposed routes to ion fragments 78
Figure 2.4 Formation of alternative m/e 123 (a) and124 (b) fragments 79
Figure 2.5 Formation of fragment m/e 151 fragments 80
Figure 2.6 Acylium ion 80
Figure 2.7 Clevenger hydro-distillation apparatus 85
Figure 2.8 Santonin 86
Figure 2.9 Santonin calibration curve 92
Figure 2.10 Santonin standard line 93
Figure 2.11 Chromatograph of the essential oil extract of A.montana flowers 96
Figure 2.12 Chromatograph of the essential oil extract of A.montana stems 97
Figure 2.13 Chromatograph of the essential oil extract of A.montana roots 98
Figure 2.14 Chromatograph of the essential oil extract of A.chamissonis flowers 99
Figure 2.15 Results of the essential oil analysis 101
Figure 2.16 Sample chromatograph of the Orcadian A. montana solvent extract 102
Figure 2.17 Sample chromatograph of the commercially available
A. montana solvent extract 102
Figure 2.18 Retention time of Peak 1 103
xii Figure 2.19 11 α13-dihydrohelenalin 103
Figure 2.20 The mass spectra of peak 1 103
Figure 2.21 Formation of the 191 fragment 104
Figure 2.22 Retention time of Peak 2 104
Figure 2.23 Helenalin 105
Figure 2.24 The mass spectra of peak 2 105
Figure 2.25 Retention time of Peak 3 106
Figure 2.26 The mass spectra of peak 3 106
Figure 2.27 The acetyl group 107
Figure 2.28 The removal of acetic acid from acetyl helenalin 107
Figure 2.29 Retention time of Peak 4 107
Figure 2.30 The mass spectra of peak 4 108
Figure 2.31 The removal of acetic acid from acetyl dihydrohelenalin 108
Figure 2.32 Retention time of Peak 5 109
Figure 2.33 The mass spectra of peak 109
Figure 2.34 The isobutyryl group 110
Figure 2.35 Retention time of Peak 6 110
Figure 2.36 The mass spectra of peak 6 111
Figure 2.37 Retention time of peak 7 111
Figure 2.38 The mass spectra of peak 7 112
Figure 2.39 The methacyrl group 112
Figure 2.40 Retention time of peak 8 112
Figure 2.41 The mass spectra of peak 8 113
Figure 2.42 Retention time of peak 9 113
Figure 2.43 The mass spectra of peak 9 114
xiii Figure 2.44 The 2-methylbutyryl group 114
Figure 2.45 The isovaleroyl group 115
Figure 2.46 Retention time of peak 10 115
Figure 2.47 The mass spectra of peak 10 115
Figure 2.48 Retention time of peak 11 116
Figure 2.49 The mass spectra of peak 11 116
Figure 2.50 Retention time of peak 12 117
Figure 2.51 The mass spectra of peak 12 117
Figure 2.52 Retention time of peak 13 118
Figure 2.53 The mass spectra of peak 13 118
Figure 2.54 The tigloyl fragment 119
Figure 2.55 Retention time of peak 14 119
Figure 2.56 The mass spectra of peak 14 120
Figure 2.57 Sesquiterpene lactone content of commercially available
& Orcadian A.montana flowers 122
Figure 2.58 Total sesquiterpene lactone content of commercially available & Orcadian A.montana flowers 123
Figure 2.59 Sesquiterpene lactones content of A.montana flowers at different stages 126
Figure 2.60 Total sesquiterpene lactone content of A.montana flowers at different stages 127
Figure 2.61 Chromatograph of the A. chamissonis solvent extract 128
Figure 2.62 Retention time of peak 15 128
Figure 2.63 The mass spectra of peak 15 129
Figure 2.64 The fragmentation of 4-O-Acetyl-6-desoxychamissonolide 130
xiv Figure 2.65 4-O-Acetyl-6-desoxychamissonolide 130
Figure 2.66 Retention time of Peak 16 131
Figure 2.67 The mass spectra of peak 16 131
Figure 2.68 Chamissonolide 132
Figure 2.69 Fragmentation of chamissonolide 132
Figure 2.70 Retention time of Peak 17 133
Figure 2.71 The mass spectra of peak 17 133
Figure 2.72 Arnifolin 134
Figure 2.73 Fragmentation of Arnifolin 134
Figure 2.74 Sesquiterpene lactone content of A.chamissonis flowers 136
Figure 2.75 Levels of sesquiterpene lactones in different stages of floral development and of the leaves and stems in A.chamissonis 138
Figure 3.1 Factors that can cause plant stress 147
Figure 3.2 Photo of some of the Arnica trials 149
Figure 3.3 Map of the Arnica trials 149
Figure 3.4 AM1 (screened) 150
Figure 3.5 AC1 (screened) 151
Figure 3.6 AM2 (not screened) 151
Figure 3.7 AC2 (not screened) 152
Figure 3.8 Number of A. montana flowers picked 153
Figure 3.9 Number of A. chamissonis flowers picked 153
Figure 3.10 Flowers of A. montana 155
Figure 3.11 Photos of the climate station at Hundland Loch 157
Figure 3.12 Map of the climate station at Hundland Loch 158
Figure 3.13 Chart showing the percentages of sand, silt and clay
xv and the textural classifications 160
Figure 3.14 Diagram of both soil pits illustrating the similar nature of the two pits 167
Figure 3.15 Detailed photos of soil pit 1 illustrating different horizons 168
Figure 3.16 Graph of temperature data from 2003-2007 173
Figure 3.17 Graph of rainfall recordings from 2003-2007 174
Figure 3.18 Graph of recorded sun from 2003-2007 174
Figure 3.19 Graph of maximum recorded wind speed and gale days from 2003-2007 175
Figure 3.20 Photo of A. chamissonis shoots 180
Figure 3.21 Photo illustrating the even cover of A. chamissonis plants. 180
Figure 3.22 Photo illustrating the clumped nature of A. montana shoots 181
Figure 3.23 Photo illustrating the rosette form of A. montana plants. 181
Figure 3.24 Slug damage on the leaves of A. chamissonis 182
Figure 3.25 Slug eating the leaves of A. chamissonis 183
Figure 3.26 Slug damage to flowers of A. chamissonis, with cuckoo spit 183
Figure 3.27 Scatter plot of minimum wind chill against dry weight for A. chamissonis and A. montana 187
Figure 3.28 Scatter plot of wind chill against flower number for
A. chamissonis and A. montana 188
Figure 3.29 Scatter plot of minimum temperature against dry weight for A. chamissonis and A. montana 189
Figure 3.30 Scatter plot of minimum temperature against flower number for A. chamissonis and A. montana 190
Figure 3.31 Scatter plot of average wind speed against dry weight
xvi for A. chamissonis and A. montana 192
Figure 3.32 Scatter plot of number of gale days against dry weight for A. chamissonis and A. montana 193
Figure 3.33 Scatter plot of average summer wind speed against dry weight for A. chamissonis and A. montana 194
Figure 3.34 Scatter plot of total rainfall against average wind speed 195
Figure 3.35 Scatter plot of total summer sun hours against average summer wind speed 195
Figure 3.36 Scatter plot of total hours of sun against dry weight for A. chamissonis and A. montana 197
Figure 3.37 Scatter plot of maximum daily sun against dry weight for for A. chamissonis and A. montana 198
Figure 3.38 Scatter plot of total hours of summer sun against dry weight for A. chamissonis and A. montana 199
Figure 3.39 Scatter plot of total summer sun hours against mean summer temperature 200
Figure 3.40 Scatter plot of mean summer temperature against dry weight for A. chamissonis and A. montana 201
Figure 3.41 Scatter plot of mean summer temperature against dry weight for A. chamissonis and A. montana 202
Figure 3.42 Scatter plot of total rainfall against dry weight for
A. chamissonis and A. montana 203
Figure 3.43 Scatter plot of maximum daily rainfall against dry weight for A. chamissonis and A. montana 204
Figure 3.44 Total sesquiterpene lactone content for AM1 and AM2
xvii over a four year period 207
Figure 3.45 Total sesquiterpene lactone content for AC1 and AC2 over a four year period 207
Figure 3.46 Individual sesquiterpene lactone content from AM1 & AM2 over a 4 year period 208
Figure 3.47 Correlation of isovaleroyl helenalin against total summer sun hours in AM2 211
Figure 3.48 Correlation of number of gale days against chamissonolide
AC1 and AC2 212
Figure 3.49 Sesquiterpene lactone production in A. montana over the 2006 field season 215
Figure 3.50 Sesquiterpene lactone production in A. montana over the 2007 field season 215
Figure 3.51 Sesquiterpene lactone production in AC1 over the 2006 field season 217
Figure 3.52 Sesquiterpene lactone production in AC2 over the 2006 field season 218
Figure 3.53 Sesquiterpene lactone production in AC1 over the 2007 field season 218
Figure 3.54 Sesquiterpene lactone production in AC2 over the 2007 field season 219
Figure 4.1 Trial AM3 237
Figure 4.2 Spread of disease in AM3 (2005) 238
Figure 4.3 Spread of disease in AM3 (2006) 238
Figure 4.4 Trial AM5 (2005) 240
xviii Figure 4.5 Trial AM5 (2006) 241
Figure 4.6 Average number of plants per plot surviving in each trial by treatment (AM3) 245
Figure 4.7 Average number of plants per plot surviving in each trial by treatment (AM5) 246
Figure 4.8 Log transformed plant survival data for AM5 in 2007 248
Figure 4.9 Yield data for AM3 in 2004 249
Figure 4.10 Average flower dry weight for AM3 in 2004 249
Figure 4.11 Number of flowers produced per plant AM5 (2006) 252
Figure 4.12 Flower dry weight produced per plant AM5 (2006) 252
Figure 4.13 Percentage content of heptacosane 254
Figure 4.14 Percentage content of peak D 256
Figure 4.15 Percentage content of perhydrofarnesylactone 256
Figure 4.16 Content (mg/ml) of sesquiterpene lactones in the solvent extract of AM5 259
Figure 4.17 Content (mg/ml) of isobutyrl helenalin in the solvent extract of AM5 259
Figure 4.18 Content (mg/ml) of isovaleroyl helenalin in the solvent extract of AM5 260
Figure 4.19 Content (mg/ml) of tigloyl dihydrohelenalin in the solvent extract of AM5 260
Figure 4.20 Total helenalin ester content (mg/ml) in the solvent extract of AM5 261
Figure 4.21 Total sesquiterpene lactone content (mg/ml) in the solvent extract of AM5 261
xix Figure 5.1 AM7 277
Figure 5.2 Photograph of AM7 (2005) 278
Figure 5.3 Photograph of ACM1 (2006) 279
Figure 5.4 Typical plot of ACM1 279
Figure 5.5 ACM1 280
Figure 5.6 Photograph of ACM2 (2006) 281
Figure 5.7 ACM2 282
Figure 5.8 AM7 in October, 2005 286
Figure 5.9 AM7 in 2006 286
Figure 5.10 AM7 in June 2006 287
Figure 5.11 Number of plants surviving, by mulching treatment 288
Figure 5.12 Photograph illustrating the lack of effectiveness of the mulches against chickweed 288
Figure 5.13 Log of number of plants surviving against bed at different levels of mulch 290
Figure 5.14 Average dry weight of A.montana flowers against bed at different levels of mulch 292
Figure 5.15 Total dry weight of A.chamissonis flowers against bed at different levels of mulch 293
Figure 5.16 Total fresh weight of A.chamissonis flowers per plot against bed at different levels of mulch 293
Figure 5.17 Total number of flowers of A.chamissonis per plot against bed at different levels of mulch 294
Figure 5.18 Acetyl helenalin in A.montana extract against different levels of mulch 297
xx Figure 5.19 Isobutyryl dihydrohelenalin in A.montana extract against different levels of mulch 297
Figure 5.20 Methacryl helenalin in A.montana extract against different levels of mulch 298
Figure 5.21 Tigloyl helenalin in A.montana extract against different levels of mulch 298
Figure 5.22 Sesquiterpene lactones in A.montana extract from different seed sources (ACM2) 300
Figure 5.23 Individual sesquiterpene lactones in A.montana extract from different seed sources (ACM2) 302
Figure 5.24 Sesquiterpene lactones in A.chamissonis extract from different seed sources (ACM2) 304
Tables
Table 1.1 The tribes of the Compositae as classified by Bentham 9
Table 1.2 Terpene classification 32
Table 1.3 The ester groups of Arnica’s sesquiterpene lactones 44
Table 1.4 A. montana dosage 58
Table 2.1 The ten editions of the German Pharmacopoeia 65
Table 2.2 Percentage content of the sesquiterpene lactones extract of A. chamissonis flowers 70
Table 2.3 The five chemotypes of A. chamissonis and their source locations 71
Table 2.4 Partial amounts of the sesquiterpene lactones in the flower heads of A. chamissonis subsp foliosa 82
xxi Table 2.5 Correction factors used for the calculation of individual sesquiterpene lactone content in A. montana flowers 90
Table 2.6 Correction factors used for the calculation of individual sesquiterpene lactone content for A. chamissonis flowers 91
Table 2.7 Level of variation in repeated analysis of 0.08 mg/ml santonin standard 93
Table 2.8 Compounds identified in the essential oil analysis 100
Table 2.9 Sesquiterpene lactone content of commercially available and Orcadian A. montana flowers 121
Table 2.10 Sesquiterpene lactone content in the different stages of floral development in A. montana 125
Table 2.11 A. chamissonis extract 135
Table 2.12 Results for A. chamissonis floral stages 137
Table 3.1 Summary harvest data for 2003 harvest of A.montana & A. chamissonis in exposed and sheltered trials 155
Table 3.2 Harvest data from the 2003 harvesting trial of A. montana 156
Table 3.3 Kirkwall monthly averages (1971-2000) 159
Table 3.4 Soil type 162
Table 3.5 Soil profile characteristics 166
Table 3.6 Extractable mineral content 169
Table 3.7 Orkney climate data 171
Table 3.8 Climate data for the May-July period of each field season 172
Table 3.9 Yield data for AM1 and AM2 177
Table 3.10 Yield data for AC1 and AC2 178
Table 3.11 Correlation coefficients for the main climate factors and
xxii yield for each plot 185
Table 3.12 Sesquiterpene lactone content (mg/ml) of AM1 and AM2 over a four year period 209
Table 3.13 Sesquiterpene lactone content (mg/ml) of AC1 and AC2 over a four year period 210
Table 3.14 Correlations of the AM2’s sesquiterpene lactone content with climate data 210
Table 3.15 Correlations between AC1 and AC2’s sesquiterpene lactone content and climate data 211
Table 3.16 Sesquiterpene lactone (SL) content of AM2 produced in
2006 213
Table 3.17 Sesquiterpene lactone (SL) content of AM2 produced in
2007 214
Table 3.18 Sesquiterpene lactone (SL) content of AC1 and AC2 produced in 2006 216
Table 3.19 Sesquiterpene lactone (SL) content of AC1 and AC2 produced in 2006 217
Table 4.1 AM3 treatments 236
Table 4.2 Yield data for AM3 (2004) 250
Table 4.3 Yield data for over flowers in AM3 (2005) 250
Table 4.4 Yield data for AM5 (2006) 251
Table 4.5 Yield of oil by treatment for AM3 253
Table 4.6 Profile of essential oil from partially open and over flowers 255
Table 4.7 ANOVA results for significance table of variation in sesquiterpene lactones 258
xxiii Table 5.1 ICRCL analysis of the mineral mulch 284
Table 5.2 XRF analysis of the mineral mulch 284
Table 5.3 Number of plants surviving in AM7 (2006) 287
Table 5.4 Number of plants surviving in ACM1 by column 289
Table 5.5 Table of F-values for sesquiterpene lactones by treatment in ACM1 ( A.montana) 296
Table 5.6 Table of F-values for sesquiterpene lactones by treatment in ACM1 ( A.chamissonis) 299
Table 5.7 Table of F-values for sesquiterpene lactones by treatment in ACM2 ( A.montana) 301
Table 5.8 Table of F-values for sesquiterpene lactones by treatment in ACM1 ( A.chamissonis) 303
Table 6.1 Factors believed to influence flower dry weight and/or flower number in Arnica 333
Table 6.2 Factors believed to influence sesquiterpene lactone production in Arnica 334
Table 6.3 Factors believed to influence mortalities in A.montana 334
xxiv List of Abbreviations
AI Agronomy Institute
AC Arnica chamissonis
Al 2O3 Aluminium Oxide
AM Arnica montana
ANOVA Analysis of variance
ARBO From ARnica and BOmme – cultivar
COD Chemical Oxygen Demand
DCM Dichloromethane
DEFRA Department for Food, Environment and Rural Affairs
EI Electron Impact
EtOAc Ethyl Acetate eV Electron Volts
GC MS Gas Chromatograph Mass Spectrometer
GDR German Democratic Republic
H2O Water
HP LC High Performance Liquid Chromatography
ICRCL Inter-Departmental Committee on the Redevelopment of
Contaminated Land
K2O Potassium Oxide
LC MS Liquid Chromatograph Mass Spectrometer
LOD Limit of Detection
LOI Loss on Ignition
LOQ Limit of Quantitation
MeOH Methanol
xxv N Nitrogen
PAH Polycyclic Aromatic Hydrocarbons
P2O5 Phosphate
SAC Scottish Agricultural College
SL Sesquiterpene Lactone
UV Ultra Violet
XRF X-Ray Fluorescence
xxvi 1 Introduction
1.1 Research context
Throughout human history, people have utilised the vast diversity of plant material available for food, fuel, perfumes, dyes and drugs. In prehistoric times, cultivators would have been restricted to the exploitation of local plants, but the modern farmer can choose to cultivate a range of potential crops, often originating from other countries or continents. The choice available to the farmer can be extensive, whilst the information available as to suitability and potential yield, is often limited.
To this end, a number of research institutes have embarked upon field trials of a wide range of non endemic species in order to determine the suitability of such crops for cultivation in their local environment. This research project aims to contribute to our understanding of one such crop in the context of the Orkney environment.
Botanical extracts are used as dietary supplements (‘nutraceuticals’), botanical drugs in the form of complex extracts (e.g. senna for constipation), and as a source for extracted drugs (e.g. taxol or morphine). Such compounds, utilised by humans to maintain their health and well being, are often synthesised by plants to maintain and defend themselves from herbivory (Picman, 1986, Harborne, 1988 and Schmidt et al. 2008).
Patients have been advised to chew willow bark to reduce inflammation for thousands of years (Flükiger & Hanbury, 1874). Since 1897, when this bark was used to derive Aspirin from salicylic acid, the role of natural products in drug
1 discovery has continued to grow (Schmidt et al. 2008). Of the small molecule drugs discovered since 1981: 37% were synthetic, 5% were natural products, 27% were derived from natural products and 31% were created synthetically but were analogues of, or inspired by, natural products (Newman & Cragg, 2006).
Additionally, despite the availability of a wide range of synthetic medical treatments and cures, the field of herbal medicine still has potential for expansion as interest in traditional herbal medicines continues to grow (Harborne & Baxter,
2001).
A significant proportion of the Asteraceae include those exploited for their extracts.
Plants such as Chamomile ( Matricaria recutita L. ), Feverfew ( Tanacetum parthenium L.), Milk thistle ( Silybum marianum L.) and Echinacea ( Echinacea purpurea L.) in particular have undergone a resurgence in demand (Harborne &
Baxter, 2001). Although the aforementioned are all members of the Asteraceae, the wide ranging applications for the extracts of this family are more the result of systematic, chemotaxonomic and pharmacological research than of tradition alone. Many compounds identified in the Asteraceae have served as models for drug synthesis and it is certain more remain to be discovered (e.g. Wagner, 1977,
Simmonds, 2003).
1.2 Arnica extract
Arnica is an example of an Asteraceae genus that is currently exploited for its extract, primarily in the form of topical treatments for bruising (Wagner et al.
2004a). Although a number of species within this genus have been used for extract production, the most popular species is Arnica montana L. (DAB 7, 1964
2 and DAB 10, 1991) (Fig. 1.1) . However, its popularity has led to it becoming overexploited in the wild (Kathe et al. 2002) and to increased interest in its cultivation (Bomme & Daniel, 1994; Douglas et al. 2004, Spitaler et al. 2006).
Figure 1.1 – Arnica montana
Arnica extract first became popular in the 1700s (Flückiger & Hanbury, 1874), as a remedy for a wide range of ailments - from rheumatism and bruises, to heart disease and fevers - and has a long history in the folk medicine of the countries in which it is indigenous. Johann Michael Fehr and several other physicians introduced A. montana into German medicine in 1712 under the name of Panacea lapsorum (Hamilton, 1852). It appeared in the London Pharmacopoeia of 1788 and by the mid 1850s Flückiger & Hanbury (1874) reported that it was a popular tincture for "preventing the blackness of bruises".
The recipes for home made Arnica ointments have been handed down from family to family. Recommended for the American home medicine cabinet in the early 20 th century (Brown, 1903), it is still recommended today (Thompson, 2009). Flückiger
3 and Hanbury in 1874 described the pharmaceutical part of the plant to be the root, although before this the flowers 1 were used. The British Pharmacopoeia instructed that the tinctures should be produced from the roots and so flower tinctures went out of use in Great Britain. A renaissance in herbal remedies originating from
Europe has had the result that currently in Britain and Europe it is now mostly the flowers that are employed in the production of both ethanol extracts and homeopathic dilutions (British Herbal Medicine Association, 1996 and Small &
Catling, 1999). This may in part be due to the cultivation difficulties experienced by many growers: the flowers are ‘renewable’, the roots are not. It is also much harder to prepare the root extract when cleaning and washing is taken into account, than it is to just use the dried flowers.
Arnica flower extract is added to a wide range of soaps, shampoos, ointments and lotions (Hausen, 1996) and used to prepare products aimed specifically at treating sports injuries (Kucera et al. 2003). Externally, it is used as a treatment for bruises and swelling where the skin is unbroken.
The main active compounds of Arnica are believed to be sesquiterpene lactones
(Willuhn, 1998) which are produced in trichomes (Heywood et al, 1977) and are common in the Asteraceae (Fischer, 1991). As well as being potent anti- inflammatory agents, these compounds have been shown to have cytotoxic and anti-tumour activity by serving as a potential telomerase inhibitor for cancer cells
(Huang et al. 2004) and an inducer of leukaemia HL-60 cell differentiation (Kim et al. 2005). In the plant, sesquiterpene lactones are believed to play a role in
1For ease of writing, the use of ‘flower’ in the text will be used in reference to the capitulum unless stated otherwise.
4 defence against insects (Guillet et al. 2000) and to act as anti-feedants (Picman,
1986). The main sesquiterpene lactones in Arnica are helenalin, dihydrohelenalin and their esters (Willuhn & Leven, 1995).
In homeopathy, Arnica extract has been used for many years to treat external injury and shock (Hamilton, 1852, Kucera et al. 2003, Oberbaum et al. 2005 and
Brinkhaus, 2006). Homeopathic tinctures are the only form in which Arnica may be taken internally due to the toxic properties of its active compounds (Anon, 2001).
In homeopathy the Arnica extract (often from the root) is diluted one part with either nine or ninety-nine parts of distilled water and / or alcohol and vigorously shaken (succussed). One part of the diluted medicine is then further diluted and the process is repeated. Dilutions of 1 to 10 are represented by the Roman numeral X (1X = 1/10) or the numeral C (1C=1/100). So an Arnica extract described as being 30X dilution will have been diluted 10 30 times (Hahnemann,
1993). It is believed by practitioners that the vigorous shaking (succussion) at each stage of dilution leaves behind some form of ‘essence’ (Hahnemann, 1993). It is non-toxic in this form as statistically, no molecules of the active compounds are likely to be present.
Chakrabarti et al. (2001) found in a blind study that sonicated mice fed with
Arnica 30X showed appreciably reduced DNA damage compared with those given a control, although it was pointed out by Bonamin & Endler (2010) that the control group was not ideal as there was no vehicle and no sucussion (i.e. the control was distilled water). Similar work by Lussignoli et al. (1999) found that in experiments on rats, oedemas healed faster when given a homeopathic remedy containing
5 Arnica and other plant extracts. Oberbaum et al. (2005) found in a double blind randomised placebo controlled clinical trial that homeopathic Arnica extract significantly reduced post partum bleeding (although all studies involved very low numbers). However, Macedo et al. (2004) found that homeopathic Arnica had to be taken pre-treatment to be effective and in a study of knee surgery, Brinkhaus et al. (2006) in a double blind, placebo controlled study found that Arnica was of significant benefit only for cruciate ligament reconstruction and not for arthroscopy or for artificial knee joint implantation.
Externally and at much higher concentrations, Arnica extract is used as a treatment for bruising and external injury. Alfredo et al. (2009), found it to be an effective anti-inflammatory on acute muscle lesions and Kucera et al. (2003) demonstrated the analgesic effect of Arnica in a single blind trial. In a double blind placebo controlled trial of topical Arnica gel on reducing post-laser treatment induced bruises, Alonso et al. (2002), found no significant difference between the cream vehicle alone, and the vehicle with Arnica , although this study made no attempt to determine the concentration of the active compound in the Arnica gel.
Meanwhile, Knuesel et al. (2002) found that an Arnica gel (Bioforce) was an effective treatment for osteoarthritis of the knee.
As for other alternative therapies, while evidence for efficacy is mixed (e.g. controls are often no treatment at all, rather than a massage with a gel that does not contain Arnica extract, e.g. Alfredo et al. 2009), there remains enough interest to ensure the continued demand for extract.
6 1.3 The market
A. montana has become overexploited in the wild due to: the increased sales of herbal medicines (Lange, 1998 and Kathe et al. 2002), the unstable yields of cultivated plants (Cassells et al. 1999), the loss of habitat due to agricultural change (Fabiszewski & Wojtun, 2001) and the high cattle density in Arnica habitats (Kathe, 2005). Trade in the species is now monitored and it is listed in
Annex D of the European Union's Council regulation No. 338/97 (Lange, 1998). As a result, A. montana cannot be harvested from the wild in Germany, Hungary or
France. However, it is still harvested from the wild in countries that list it as vulnerable, such as Romania and in regions where it is critically endangered such as the Balkans (Lange, 1998 and DAB 10, 1991). In 2001, 28 tonnes of A. montana flowers were collected from Transylvania alone with 90% destined for the
International market (Kathe et al. 2003). Despite conservation concerns, Germany alone markets at least 300 drug preparations containing cultivated Arnica extract
(Small & Catling, 1999) and strong market demand is still evident (Burnie, unpublished).
As most Arnica is hand harvested from the wild, the countries where this material is collected tend to have low labour costs (Kathe, 2003). It has been estimated that
50 tonnes of dried flowers are used per year from both cultivated and wild sources
(Lange, 1998). The average weight of a fresh flower is 0.90g whilst a dried flower is 0.21g (Martin, unpublished), so 50 tonnes of dried flowers would be roughly equivalent to 214 tonnes of fresh flower material. Approximately 4,760 flower heads would have to be collected to obtain one kilogramme of dried Arnica flowers and 238 million would be required for 50 tonnes dried Arnica flowers.
7
Prices paid for a kilogram of dried flowers have varied from the equivalent of £8 per kg in 2001 to approx £51 in 2002 (Kathe et al. 2003). It was available for £30 / kg in Germany in 2005 (Kathe, 2005) and £44 / kg in the UK (Organic Herb
Trading Co. 2006) whilst a kilogramme of root fetched up to £60 / kg (Kathe,
2005). With an estimated market demand for the flowers of approximately 50 tonnes per year (Lange, 1998), and a current estimated market value of approximately £20 per kilogram (Burnie, unpublished), this equates to an approximate total market value of £1,000,000. However, as a proportion of the harvested Arnica is wild harvested, these figures may be subject to change with increased protective legislation.
To meet demand, attempts have been made to find potential alternative species which are easier to grow and harvest, but also have acceptable levels of the relevant active compounds. A species identified as being similar to A. montana in terms of active compounds is Arnica chamissonis Less. ssp. Foliosa (Nutt.)
Maguire (1943). Studies by Leven & Willuhn, (1987), Willuhn & Kresken (1981),
Willuhn et al. (1983) and Belin et al. (1998) have all found that A. chamissonis extract contains a similar, if not identical, range of sesquiterpene lactones. As a result, it has now been accepted as a substitute for A. montana extract by the
Pharmacopoeias of the German Democratic Republic and the Soviet Union
(Schröder & Merfort, 1991 and Willuhn et al.1994).
8 1.4 The Taxonomy of Arnica
As described earlier, Arnica belongs to the family Asteraceae (previously
Compositae). It is one of the largest plant families and one in which there is considerable uniformity in the floral and fruit structures (Heywood, 1971). Initially, the Asteraceae (or Compositae as they were known as then), were classified into
13 tribes (Table 1.1) by Bentham in 1873.
Table 1.1 - The tribes of the Compositae as classified by Bentham (1873)
Name (original spelling in Tribe parentheses) 1 Veronieae (Vernoniaceae) 2 Eupatorieae (Eupatoriaceae) 3 Astereae (Asteroideae) 4 Inuleae (Inuloideae) 5 Heliantheae (Helianthoideae) 6 Helenieae (Helianthoideae) 7 Anthemideae 8 Senecioneae (Senecionideae) 9 Calenduleae (Calendulaceae) 10 Arctotideae 11 Cynareae (Cynaroideae) 12 Mutisieae (Mutisaceae) 13 Cichorieae (Cichoriaceae)
However, this was subject to much debate (due mainly to the uniformity of features), until investigations revealed that the phylogenetically young family
(Herout, 1971) produced a large variety of secondary metabolites and were unlike any other in their range of chemical constituents (Heywood et al. 1977). In an overview of the secondary metabolites found in Arnica species, Willuhn et al.
(1991) suggested that in terms of evolutionary development, Arnica should be placed in or near the Heliantheae with a close affinity to the Gaillardiinae.
9 However, they also stressed the importance of further investigations in this area to define the complex chemistry involved. This is discussed further in section 1.9.3.3.
Figure 1.2 shows the current classification of the Angiosperms from the
Angiosperm Phylogeny Group (APG, 1998). The Asteraceae sit within the
Asterales, in the Euasterids II. Most species of the Asterales can be classified as
Asteraceae which is generally acknowledged as one of the most advanced families of flowering plant and contains over 20,000 species (Ingrouille & Eddie,
2006).
All species within the Asteraceae have a flower head receptacle containing many florets, which is known as a capitulum. This provides a large target for pollinators and protects the ovule and the seed. The florets themselves are tubular, have a single ovule and can vary in number depending upon the species. The fruit
(achene) of the family normally has a crown of pappus which can also be highly variable. The Asteraceae have enjoyed particular evolutionary success due to their high levels of genetic diversity. They are multiallelic, homomorphic, sporophytic and self incompatible, but display flexibility in that many of the weed species are secondarily self compatible and self pollinating which has contributed in no small part to the world wide distribution of this family (Ingrouille & Eddie,
2006).
10
Figure 1.2 - Phylogenetic interrelationships of the orders of flowering plants
APG (1998) & APG (2002)
11 Arnica is a circumboreal composite genus composed of five subgenera: Arctica,
Austromontana, Chamissonis, Arnica, and Andropurpurea and about thirty-two generally well-segregated species mostly found in North America (Maguire, 1943).
A. montana was first described by Linnaeus (1753) whereas A. chamissonis was described by Lessing (1831).
1.4.1 Arnica montana
1.4.1.1 Distribution and growth
Although A. montana is commonly referred to as Arnica , common names include mountain Arnica , mountain daisy, mountain tobacco, mountain snuff, leopard’s bane, sneeze-wort and fall-kraut (Small & Catling, 1999). A. montana is found in the mountains of Europe (Hamilton, 1852) where it is native to high alpine slopes of western and central regions and on the lower plains of colder countries with poor acidic soils (Maguire, 1943, Pegtel, 1994) (Fig. 1.3). It is rarely found in lowland bogs and not in grasslands of lower elevations (Bruelheide & Scheidel,
1999). Although A. montana is normally only found at altitudes of between 1000m to 2,500m, it can be cultivated in a mixture of peat, loam and sand if sown in very early spring (Palaiseul, 1976) but is more commonly found on acidic (pH 5-6), well- drained, nutrient poor sandy podzols with a surface layer of mor humus (Pegtel,
1994). Tosco (1978), reports that A. montana can also be found in marshy meadows.
12
Figure 1.3 – Occurrence of A. montana
Catalogue of Life: 2007 Annual Checklist: The Integrated Taxonomic Information System (GBIF, 2008)
Calcium levels should be low, less than 1%, as a high concentration of calcium reduces growth and suppresses accumulation of potassium and magnesium
(Raison et al. 2000, Jenelten & Feller, 1992). It has also been suggested that light disturbance of the topsoil layer (weeding, turf cutting etc) favours growth by enhancing its competitiveness against grasses such as Deschampsia flexuosa L.
(Pegtel, 1994).
A. montana is sensitive to the atmospheric deposition of nitrogen and sulphur dioxide (Dueck & Elderson, 1992), which are acidifying pollutants, often leading to changes in both soil and water properties (Hall et al. 2006). A recent investigation into the addition of different ratios of reduced (alkali) and oxidised (acidic) nitrogen on heathland species including A. montana (van den Berg et al. 2008), demonstrated the mortality of A. montana to be unaffected by soil acidity when the aluminium content of the soil was low. In their trials, the biggest increase in biomass was found under the liming treatment. This is likely to be due to soils with
13 a high pH being unlikely to leach aluminium ions and hence decreasing the probability of aluminium toxicity (Scheffer & Schachtschabel, 1979).
A. montana reproduces both sexually and asexually although it is largely self- incompatible (Luijten et al. 1996). Clonal reproduction is via rhizomes (Kahmen &
Poschlod, 2000) and one plant can produce several new stems via rhizomes each year (Schwabe, 1990). Genotypic analysis in the Netherlands has revealed that dense clusters often have identical genotypes (Luijten et al. 1996) although this was not found to the same extent in the Rhön, Germany (Kahmen & Poschlod,
2000).
1.4.1.2 Description
A. montana is a deciduous, rosette, clump forming, rhizomatous and polycarpic perennial with large bright yellow/orange aromatic composite flowers on erect unbranched stems (Maguire, 1943 and Flückiger & Hanbury, 1874) (Fig. 1.4 & Fig
1.5).
Figure 1.4 – A. montana flowers
14 The leaves are pale green, ovate, basal and simple, often widest beyond the middle. From the centre of the rosette, the flower stems rise and can grow up to 50 cm in height. The flower capitulum is about 2 cm in diameter and about 1.5 cm deep; the peduncle is 2-3 cm in length. There are many receptacles, separate, 3-8 mm in diameter, flat or slightly convex with a honeycomb appearance with two rows of about 20-25 dark green, linear- lanceolate pubescent involucral bracts
(British Herbal Medicine Association, 1996). The disc bears a large number of florets, which are about 15 mm long. There are about 20 peripheral ligulate florets approximately 20 mm long with dark yellow pistellate rays. The ovary is hairy, 4-8 mm long and is crowned with a pappus of white filaments 4-8 mm in length
(Pharmacopoeia, 2002) (Fig.1.6). Strykstra et al. (1998) found a positive correlation in A. montana between achene and pappus weight and germination and seedling quality. As heavier achenes stay closer to the point of release than lighter ones, it seems A. montana has a moderate dispersal strategy of only a few metres.
Figure 1.5 - Floral diagram of A. montana
The root is dark brown, contorted, 8-10 cm in length and on the underside there are many wiry roots, 8-10 cm long (Flückiger & Hanbury, 1874).
15
Figure 1.6 – Achenes and pappus of A. montana
Tracey Slotta @ USDA-NRCS PLANTS Database
1.4.2 Arnica chamissonis
1.4.2.1 Distribution and growth
A. chamissonis is also known as leafy Arnica and is more widely distributed than
A. montana. It flowers from June to August and can be found from France to
Russia, Alaska, through Canada and the Rockies to California and New Mexico
(Maguire, 1943, Camp & Gilly, 1943, Rickett, 1973). Over this considerable range it displays polymorphy (Camp & Gilly, 1943) which was also found by Willuhn et al.
(1994). A. chamissonis is found mainly in the US and immediately adjacent areas, extending far northwards into the interior lowlands of Canada (Fig 1.7).
16
Figure 1.7 – US distribution of A. chamissonis ssp. foliosa
Image adapted from USDA (2010)
A. chamissonis can be found on mountains and in valleys in meadows, wet places
(Cronquist, 1955 and Rickett, 1973) and alkaline soil (Plantlife, 2004). Where it grows in wet places it often has longer internodes (Camp & Gilly, 1943). A. chamissonis is less susceptible to high levels of calcium than A.montana but is sensitive to high levels of phosphate (Jenelten & Feller, 1992). It produces a large number of shoots from its root system, which allows it to compete well with weeds
(Martin, 2003).
1.4.2.2 Description
Arnica chamissonis Less. ssp. Foliosa (Fig.1.8) is an erect perennial, between 20-
100 cm tall, with solitary leafy stems that are covered with soft wavy hairs. The leaves are simple, denticulate to more often entire, lancelolate (pointed at both ends), without stalks and with hairs at the base of the involucre. They are opposite, 5-30 cm long and 1-4 cm wide, small toothed to more often entire with 5 to 10 pairs on a stem (Rickett, 1973 & Cronquist, 1955).
17
The flower heads are yellow, radiate or discoid, 3- 5 cm across with 10-16 pistillate rays, and grow in branched clusters of one to several flower heads. The involucral bracts are herbaceous, subequal and connivent (converged and touching but not fused) approximately 1 cm deep, either blunt or soft pointed. The receptacle (tip of a flower stalk bearing the floral organs) is convex and naked. The disk flowers are perfect, fertile and yellow and the style branches are flattened, truncate or very shortly appendiculate. The achenes (Fig.1.9) are cylindric or nearly so, with 5 to 10 nerved pappus of numerous straw coloured to whitish, barbellate to subplumose capillary bristles (Cronquist, 1955) and are much longer than those of A. montana
(personal observation and Fig 1.9). The rhizomes can be up to 30 cm long are unbranched and nearly naked (Camp & Gilly, 1943 and Cronquist, 1955).
Figure 1.8 –A. chamissonis flowers
18
Figure 1.9 – Achenes and pappus of A. chamissonis
Steve Hurst @ USDA-NRCS PLANTS Database
1.5 Pests and grazing
Two phytophagous insects are specific to A. montana : Tephritis arnicae L.
(Diptera, Tephritidae) and Melanagromyza arnicarum Her . (Diptera, Agromyzidae).
The eggs of T. arnicae are laid on tubular flowers which are not yet open. They develop within the flower heads and become active in May. The larvae obtain nourishment from the flowers, destroying the ovary, and when fully grown (c.a.
July) pupate inside the flower head before emerging as adults 15 days later.
Flowers that are infested with T. arnicae require winnowing or sieving after the flowers are dried (Scaltriti, 1985). Scheidel et al. (2003) found increasing damage by T.arnicae with altitude, although this could also be the result of a corresponding increase in the population density of A. montana .
19 The eggs of M. arnicarum are laid in the stem near the flower head. The larvae bore down the stem in a basal direction and when they reach the rhizome (c.a
July) they pupate and over-winter before swarming in the following April (Scaltriti,
1985). M. arnicarum is far less damaging than T. arnicae and the presence of one or two pupae in the Arnica rhizome does not alter its properties.
To date, neither species of insect has been found within the Orkney populations.
Philaenus spumarius L. (Froghopper) is common, particularly on the A. chamissonis, but is not thought to be deleterious to growth. A. montana is frequently visited by both butterflies and flies (Kahmen & Poschlod, 2000 and personal observations).
It has been noted that in areas with high numbers of rabbits A. montana plants are left untouched (personal observation). The active compounds of Arnica are often associated with a bitter taste (DAB 10, 1991) and it is common in other Asteraceae for this to be a deterrent to grazing (Heywood et al. 1977, Picman, 1986).
MØlgaard (1986) found that other plants containing constituents repellent to mammals, were readily accepted by land molluscs and A.montana is not an exception. It has been found by others to be heavily grazed by certain species of molluscs due to the combination of slow growth and ‘highly palatable leaves’
(Scheidel & Bruelheide, 1999). This suggests that in regions where A. montana dominates, they could be susceptible to heavy attack.
20 1.6 Conservation and cultivation
The value of A. montana flowers is high which has encouraged wild harvesting.
The potential for profit, combined with the loss of available habitat due to expanding agriculture has meant the number of countries listing A. montana as vulnerable, rare or endangered has increased. It is strictly protected in France,
Germany, Hungary and parts of Switzerland and has been listed in Annex D of the
European Union’s Council regulation No 338/97 (Lange, 1998 and Ellenberger,
1998). A. montana has also been listed in several Red Data Books: ‘Indeterminate’ in Kaliningrad, Ukraine; ‘Rare’ in The Czech Republic; ‘Vulnerable’ in Bosnia-
Herzegovina, Lithuania, The Netherlands, Portugal and Romania: and
‘Endangered’ in Germany and Hungary (Kathe et al. 2002).
Collecting A. montana flowers in Spain is not restricted and this is the main source of wild material imported into Germany. However, it is felt that the populations of
Arnica are decreasing in size to the extent that this will soon pose a conservation problem (Kathe et al. 2002). A permit for collecting wild flowers is required in
Romania but regardless, over 1000 kg of A. montana plant material was exported from Romania to Germany in the 1990's (Lange, 1998). There are no current limitations on the wild harvesting of A. chamissonis .
Although Kahmen & Poschlod (2000), found no sign of genetic erosion due to reduced population size in their studies of A. montana, Luijten et al. (2000) found fitness to be reduced in small populations in the Netherlands. It is hence important that a successful method of cultivation is found.
21 There are currently several countries in which A. montana is grown for commercial purposes. Although herbal growers are often fairly reticent about sharing information, herbalists often state country of origin on their products. A survey of these listed material sourced from countries such as France, Poland, Italy,
Germany, Northern India, China and Mexico (Burnie, unpublished). However, the
Mexican product is most likely to be the antimicrobial herb Heterotheca inuloides
Cass. This is sometimes referred to as Arnica (Small & Catling, 1999) but is not related. The most common source of the wild harvested flowers is Yugoslavia,
Spain, Italy and Switzerland (DAB 10, 1991). Processed Arnica extract is most likely to show France or Germany as its country of origin. (Burnie, unpublished).
Yields of Arnica vary between sources. Bezzi and Ghidini (1989) reported yields of
94 to 284 kg per hectare in Italy, although trials were affected by T. arnicae .
Delabays and Mange (1991) reported 640 kg per hectare and trials in both
Germany (Bomme et al. 1994) and Finland (Galambosi et al. 1998) have demonstrated yields over 1000 kg/ha in the third year. However, multiannual, multilocational field trials in France, Germany and Switzerland have either not shown yield stability (Cassells et al. 1999) or have not proved profitable (Lange,
1998). The former is often the result of difficulty with the germination of seeds, the frequent loss of plants due to disease or strong chlorosis, generally poor plant development or the feeble setting of flowers (Raison et al. 2000), while the latter is likely due to a combination of unstable yields and the labour cost involved in hand harvesting. It has been estimated that 6,000 hours of labour would be required to hand harvest an eventual yield of 1000 kg of dried flowers (Burnie, unpublished).
22 Douglas et al. (2004) simulated the mechanical harvest of A. montana and demonstrated that the quality of the extract can be above the minimum quality level as set by Willuhn (1998) (see section 1.9.3.3). However, in an attempt to further maximise yield, selective breeding programmes are underway in New
Zealand to increase uniformity of plant structure, time of flowering, flower yield and flower quality (Douglas et al. 2004).
Attempts have also been made to breed a cultivar of Arnica that is adapted to plantation conditions and Bomme and Daniel (1994) have demonstrated that, via selective breeding, yields of A. montana can be increased to 1330 kg/ha. Indeed, it may be that the variety ARBO (AR from Arnica and BO from Bomme) has a high and consistent enough yield to justify the removal of Arnica chamissonis from the herbal pharmacopoeia. This claim is yet to be confirmed via published trial results although it has been shown that levels of sesquiterpene lactones did not vary with altitude in Austrian plantations of ARBO, in plots between 590 and 2230m (Spitaler et al. 2006).
Attempts have also been made to propagate both A. montana and A. chamissonis by alternative methods such as micropropagation (Buthuc-Keul & Deliu, 2001 and
Cassells et al. 1999 respectively). Although successful, these methods would significantly increase the costs of production and would only be feasible as a large scale source of material. Initial trials have shown A. chamissonis to grow vigorously in Orkney with little or no attention (Martin & French, unpublished), so micropropagation is unlikely to be required in this case. However there may be
23 other advantages associated with micropropagation of clonal Arnica , particularly in terms of improving quality.
1.7 Alternative agricultural crops and the Highlands and Islands
Mainstream agriculture can be defined as meat and cereal production (Thirsk,
2000), whilst alternative agriculture is essentially the production of all other crops.
Although normally perceived as a modern development, there have actually been at least three previous phases of alternative agriculture in Britain each linked with a fall in cereal prices (Thirsk, 2000). The first wave of alternative agriculture was after the Black Death (ca 1350), when the population dropped by up to 45%
(Goldberg, 1996). The second was the overproduction following the English Civil
War in the mid 1600s, and the third was in 1879 after the agricultural depression.
Each of these waves of alternative agriculture differs in terms of details, but common themes include a breaking up of the larger farms facilitating increasing diversification and a rise in vegetarianism and herbalism (Thirsk, 2000). In 1990, the now renowned grain mountain and falling grain prices led to the Common
Agricultural Policy on set aside land (Scottish Office, 1992) which in turn led to the current phase of diversification. The wide range of crops produced in the UK
(Agriculture UK, 2009) indicates that the country is currently within another phase of alternative agriculture, though whether it is at the beginning, in the middle or coming towards the end remains to be seen.
In terms of the physical environment, the Highlands and Islands is a very diverse region in terms of both geography and climate. This diversity however, makes it an ideal location for alternative agriculture. Whilst the fertility of soils of the Moray
24 Firth, Argyll, Speyside and Orkney regions (based on Old Red Sandstone) have led to these areas being heavily farmed, the rest of the Highlands has much reduced agricultural capacity (Caird, 1972). The resulting wide range of land cover in the Highlands has created an apparent East / West divide as illustrated in figure
1.10. Consequentially, there are a range of agricultural practices in the Highlands, ranging from large scale farms in the East and on some of the Islands, to small holdings and crofts on the West Coast. Both types of farming have different requirements and suit different crops and so any study assessing the potential of an alternative crop has to consider a broad range of factors. Whilst a crop may suit a less fertile soil on a more exposed site and hence remote areas on the west could be indicated, in the case of Arnica the technology for harvesting and drying must also be readily available to the grower, which would make a more central location preferable.
Although the planting of cereal crops has declined over the past few years, with a corresponding decrease in the use of artificial fertiliser (Agriculture UK, 2009), many farmers still focus on established crops that require the nitrogen level to be enhanced. Fertilisers are widely employed and in some intensively farmed areas of the Highlands, the nitrate applied to the crops can run off into nearby streams and rivers leading to health and environmental problems (DEFRA, 2002 & Scottish
Executive, 2003). This occurs mainly in areas that are intensively farmed and tend to be producing conventional agricultural crops. In contrast, crofters (generally based on the outer zone of the Highlands (Caird, 1972) tend to use low levels of fertilisers, herbicides and pesticides (Scottish Crofting Foundation, 2009) and are
25 supported in their attempts to introduce new diversification initiatives on the land
(Crofters Commission, 2007).
Figure 1.10 - Scottish Land Cover Map, 2000
www.ceh.ac.uk/sections/seo/lcm2000_home.html
For either A. montana or A. chamissonis to be considered for such support it first needs to be determined whether Arnica will grow well in the Highlands and Islands and whether the climate conditions are conducive to high yield. Such a study would require a windy and wet environment typical of the Highlands, but also the
26 soil and resources that would allow trials of a variety of agronomic regimes and practices. To this end, trials were conducted in Kirkwall, Orkney (Fig. 1.11).
The Orkney Islands
Figure 1.11 – Map of the Highlands with the Orkney Islands highlighted
The Orkney archipelago lies 11 km north of the Scottish mainland. It consists of around 70 islands which cover 85 km from north to south and 37 km from east to west. The coastline is 800 km long, incorporating an area of over 950 square kilometres. The landscape ranges from fertile farmland to rocky cliffs, reaching a maximum height of 470m on the Island of Hoy. At 59 oN, the islands have the same latitude as St Petersburg and the Southern mainland of Greenland which means the islands enjoy a particularly long day length in the summer, albeit for a shorter season. The Gulf Stream causes the maritime island climate to be mild, but changeable and strong winds are common with an average of 52 hours of gales
27 recorded annually (Towrie, 2008). High winds limit some of the crops that can be grown in Orkney, particularly when the short growing season is taken into account, as wind flattened crops are very difficult to harvest. Although the longer day length and short growing season may mean that Orkney is a suitable site for some alpine crops, the milder climate may limit the success of those that require very cold winters to stimulate flowering.
The viability of a crop however is also dependent upon whether more profitable alternatives can be grown on the land with less input. In Orkney, the majority of the land is used for grazing. There are approximately 1000 holdings on Orkney, grazing in excess of 92,000 cattle (Scottish Government, 2003). Despite this, agricultural diversification and interest in the growth of alternative crops is increasing (Martin, unpublished).
To this end, The Agronomy Institute (AI) at Orkney College UHI was opened in
2002. All Arnica plants used in this study were grown there as the Institute had the facilities and the space for controlled field trials, as well as an ideal location. The
AI pursues research that relates to high value natural products, for example indigenous and heritage crops and plants with medical and pharmaceutical importance, as well as lower value biomass and biofuel crops. The Arnica species chosen fall into the former category and so the AI initiated field trials in 2002 and
2003 to establish the suitability of A. montana and A. chamissonis in Orkney.
Results of these trials demonstrated that both grew well during 2002 and 2003 and it was hypothesised that this might be due to the cool temperatures and long day lengths in Orkney (French, personal communication). Research for this PhD
28 programme, continued these trials by investigating the yield resulting from varied agronomic regimes, species and seed sources. However, as the quality of the floral extract is as important as the quantity, the levels of the active compounds present were also determined.
1.8 The chemistry of Arnica
In the last sixty years, the field of chemical ecology has evolved and expanded as a result of improved chromatographic and spectroscopic techniques (Harborne,
2000). Hundreds of compounds that were previously thought to be waste products have been revealed to have important ecological roles (eg. defence against herbivores) and / or a biomedical use (Picman, 1986). Such compounds, not produced via primary biochemical pathways essential for normal growth, are commonly referred to as ‘secondary compounds’ or ‘secondary metabolites’ and in plants have a variety of roles throughout the plant’s life cycle. For example, they can take the role of anti-feedants in the plant’s defence system (e.g phytoalexins,
Essenberg, 2001) or can attract pollinators for reproduction (e.g. limonene,
Harborne, 1988, Verpoorte & Memelink, 2002). The aforementioned advances in analytical chemistry led to an exponential growth in the identification of such compounds and the total number of secondary metabolites grew from 5000 in
1950, to over 100,000 in 1994 (Buckingham, 1994).
Secondary metabolites are important for a number of reasons, but mainly because they determine the human use of the plant or plant product. For example, a berry rich in antioxidants may be used as a dietary supplement. It might also be that they contain a metabolite that either in itself or as an analogue could provide a
29 new medicinal drug. In addition, if one member of that plant family already provides a useful chemical metabolite, then it is possible that similar compounds will be found in other members of that family. Such chemical profiling can often lead to the determination of family groups themselves, particularly in genera that might initially be hard to distinguish. This field, also known as chemosystematics or chemotaxonomy will be expanded upon in section 1.8.3.2.
An understanding of secondary metabolism is also important in the field of agriculture. For example, in the mid 1970s, orange groves ( Citrus sinensis (L)
Osbeck) were attacked by the anise swallowtail butterfly, a species which is normally attracted to fennel ( Foeniculum vulagare. Mill). This was later determined to be due to both plant species emitting the same three monoterpenes; anethole, methylchavicol and anisaldehyde, which are larval feeding attractants (Harborne,
1988). This demonstrates that even for established crops, it is important that we understand what attracts insects, in order to minimise the risk of attack.
Investigations into the chemical constituents of plants can often also explain ecological relationships and lead us to a better understanding of the interdynamics of ecosystems. For example, secondary metabolites have been demonstrated to exert an effect on the growth of other plants. The desert plant Encelia farinosa
Gray ex Torr. of the Compositae inhibits other plants from growing within one metre, thereby protecting its supply of water (Harborne, 1988).
It has been estimated that only about 10% of flowering plants have been analysed for secondary metabolites and of those that have been analysed, the focus has
30 tended to be on just the key compounds (Harborne, 2000). The plants studied are often those from tropical regions, which is natural given the rapid destruction of this habitat, but many other environments remain poorly profiled as a result.
This study will not attempt to address the imbalance of tropical vs. temperate plant profiles, as the key compounds have already been identified in Arnica , but it will further investigate the production of these metabolites and how they vary by agronomic regime.
1.8.1 Terpenes
Terpenes are widespread in nature and can be found mostly in the essential oils of plants. They are compounds that are constructed from isoprene (2-methylbuta-1,3- diene) units (Ruzicka, 1953) (Fig 1.12) although isoprene units can also be found as components of other natural products.
Figure 1.12 – Isoprene unit
The terpenes are classified by how many isoprene units they contain (Table 1.2)
(Devon & Scott, 1972).
31
Table 1.2. Terpene classification
Number of Class of terpene Carbons
C10 Monoterpenes
C15 Sesquiterpenes
C20 Diterpenes
C25 Sesterterpenes
C30 Triterpenes
C40 Carotenes
This is the basis of the isoprene rule which states that the carbon skeleton of the terpenes is composed of isoprene units linked together in regular or irregular arrangement (Ruzicka, 1953) (Fig 1.13).
Regular arrangement – head to tail
Example of irregular arrangement – head to head
Figure 1.13 – Regular and irregular arrangement of isoprene units
Adapted from Ruzicka, 1953 P357
32
The simplest terpene contains two isoprene units and is referred to as a monoterpene (isoprene alone is described as a hemiterpene). The prefix ‘sesqui’ means ‘half as much again’ and is the term used to identify terpenes that are one and a half times the molecular weight of the monoterpenes (i.e. C 15 ).
1.8.2 Sesquiterpenes
The sesquiterpenes are heavier than the monoterpenes and are common in essential oils where they often play a part in the higher boiling point range of the odour profile (Williams, 1996). The sesquiterpenes display a greater natural variety than the monoterpenes (ca 7000 compared to ca 1000 of the latter, Connolly &
Hill, 1991) due to the increased structural complexity possible with an additional five carbons.
Ruzicka (1953) first put forward the theory that the carbon backbone of sesquiterpenes were synthesised via the initial formation of farnesyl pyrophosphate in the mevalonate pathway (mevalonate, isopentyl pyrophosphate, farnesyl pyro-phosphate), a hypothesis which was later confirmed by Nes &
Maclean in 1977. The steps involved in the synthesis of monoterpene, sesquiterpene and diterpene units are outlined in Figure 1.14.
33
Figure 1.14 – Organisation of terpene biosynthesis in plants.
Taken from Bohlmann et al. 1998 P4127
1.8.3 Sesquiterpene lactones
Sesquiterpene lactones are a large group of biologically active plant constituents that are common in the Asteraceae (Fisher, 1991). They have also been widely reported in the Acanthaceae, Anacardiaceae, Menispermaceae, Rutaceae,
Winteraceae and the Hepatidae (liverworts), but the greatest number (over 3,000) are from the Asteraceae ( Ibid .). They are mostly found in trichomes on the leaves and flowering heads of the Asteraceae where they can constitute up to 5 % of the dry weight (Heywood et al. 1977 and Picman, 1986). Trichomes are a major site for biosynthesis and often have the function of protecting the plant against insect predation (Wagner et al. 2004).
34 Sesquiterpene lactones such as santonin have been known for over 100 years, but interest in their structure and variety really started to increase fifty years ago when they became used both in chemotherapy (Lee et al. 1977a), and as markers in chemotaxonomy (Herout, 1971, Heywood et al. 1977, and see review by
Harborne, 2000). The latter is expanded upon in section 1.9.3.2.
The greatest number of sesquiterpene lactones are of the γ-lactone type. The formation of these involves the oxidation of one or more of the methyl groups in the isopropyl head of the farnesyl precursor to a carboxyl, then the oxidation of an adjacent methylene group to a secondary alcohol, followed by ring closure (Herz,
1977a).
1.8.3.1 Classification of sesquiterpene lactones
There are four main types of sesquiterpene lactones in the Asteraceae, the xanthanolides (e.g.xanthathin), guaianolides (e.g. geigerin), eudesmanolides (e.g. santonin), and germanacranolides (e.g. pyrethrosin). The latter group is the most common with over 1300 complex, germacrene derived sesquiterpene lactones identified (Seaman and Funk, 1983).
Sesquiterpene lactones are classified first by their carbocyclic skeleton and are then suffixed ‘olide’ by their lactonic function. A lactone is a cyclic ester (the product of condensation between an alkanol and an alkanoic acid group). In sesquiterpene lactones formed by enzymatic (and hence genetically controlled), oxidation of the pyrophosphate esters of farnesol, the lactone is an α-methylene γ- lactone moiety or a derivative (Figure 1.15). The term ‘ α-methylene’ refers to the
35 methylene group on the first carbon within the ring adjacent to the carbonyl and a
“γ-lactone” refers to a 5 membered cyclic ester. These are the most common sesquiterpene lactones (Herz, 1977b) and are classed as germacranolides
(Fischer, 1991).
O O
Figure 1.15 –The α-methylene-γ-lactone ring (germacranolides)
The sesquiterpene lactones of interest in this study are pseudoguaianolides and are of the helenanolide type where the C10 methyl group is α (Fig. 1.16).
H H
O O
Figure 1.16 – The β methyl of the ambrosanolide precursor (left) and α methyl
group of the helanolide precursor
Fisher, 1978
Biogenic pathways from the germacranolides lead to the production of eudesmanolides, elemanolides, pseudoguaianolides, seco-pseudoguaianolides, guaianolides, xanthanoilides and eudesmanoloides (Fig 1.17), with the helenanolides synthesised from the guaianolides (Willuhn & Leven, 1991). The biosynthesis of helenanolide by the acid induced cyclization of melampolide
36 epoxide was first suggested by Herz (1968) with the alternative germacrolide epoxide precursor pathway suggested by Fischer in 1978, when he identified that helenanolides (including helenalin), contain a cyclopentenone ring which would be a logical precursor for C2 hydoxylated germacranolides.
However, the variability in sesquiterpene lactones among the Asteraceae is not just due to variation in the carbon skeleton, but also to oxidative changes (among others), in the sesquiterpene lactone precursors. These changes can create side chains of hydroxyl and carbonyl groups as well as generate oxide bridges and additional carboxyl or lactonic groups. The hydroxyl groups are often then esterified and so sesquiterpene lactones often appear as acetates. Such ester groups normally present strong mass spectral peaks and are useful for identification purposes as the mass fragments of the germacranolides vary considerably and are normally limited to the parent peak and the side chains
(Merfort, 2002).
37
O
PP
Germacradiene 2E,6E-Farnesyl-OPP
O
O Eudesmanolide O
O Germacranolide O
O Guaianolide
O
O Elemanolide
O O
O
O Pseudoguaianolide O Xanthanolide O
Seco-pseudoguaianolide
Figure 1.17 – Hypothetical synthesis routes of different sesquiterpene lactones
After Fisher 1991
38 1.8.3.2 Sesquiterpene lactones and chemosystematics
Arnica belongs to the family Asteraceae, a family in which there is considerable uniformity in the floral and fruit structures (Heywood, 1971). As a result, classification within the Asteraceae, as for many other ‘natural’ families such as the Umbelliferae, is difficult and is often based upon variation in relatively insignificant characteristics (Heywood, 1971). “Anybody can recognise the family, but the genera are much more difficult” (Cronquist, 1985 P6). However, the phylogenetically young Asteraceae (Herout, 1971), produce a large variety of secondary metabolites, e.g terpenoids, flavones and flavonoids, which serve as useful taxonomic tools (Seaman and Funk, 1983). In particular, the sesquiterpene lactones serve as ideal markers compounds as they occur in great numbers in the
Asteraceae and are only found to a limited extent outside the family (e.g. parthenolide in the more primitive Magnoliaceae (Herout & Šorm, 1969)).
In 1989, Gottlieb put forward the hypothesis that the limited distribution of oxygenated terpenes (e.g the iridoids and the sesquiterpene lactones) in the more advanced plant families (e.g the Labiatae and the Asteraceae) were due to historic rising oxygen levels; the increased availability of oxygen leading to the decreased utilisation of the shikimate pathway and the enhancement of the mevalonate pathway. Gottieb (1989), proposed that the reason for this was that the higher the oxidation level of a secondary compound, the less energy required to transform it back to its initial form i.e. it will be a better antioxidant and the plant can more easily free up nutrients during senescence. However, further research into the synthetic pathways needs to be completed before this hypothesis can be accepted with any certainty.
39
When classifying species by their chemical content, the homology of biogenetically based characters must be confirmed via the discovery of shared biosynthetic pathways or routes (Seaman & Funk, 1983). Such confirmation can come from the radioactive labelling of components of the pathway, by the laboratory synthesis of the product from the hypothetical precursor, by the construction of a hypothetical pathway and the subsequent identification of intermediates within the subject organism (Geissman, 1973) or by stable isotope analysis (Akhila et al. 1987 and
Sy et al. 2001). More recently, over expression or down regulation of genes encoding biosynthetic enzymes and transcription factors have provided insights into a wide range of secondary metabolite pathways (Petersen, 2007), although to date, Arnica has not been investigated in this manner .
A framework for cladistic analysis of the Asteraceae by their sesquiterpene lactone content was first suggested by Herout 1971 (Fig 1.18), and then expanded on by
Herz (1977b). They classified the sesquiterpene lactones by the number of modifications from the compound’s original farnesyl pyrophosphate precursor and put them into one of four columns. Although grouping by complexity is essentially an artificial system, Herz argued that the more simple skeletal structures were more common throughout the Asteraceae than the relatively more complex versions, indicating a possible ancestral form. He classified the sesquiterpene lactones first by skeletal types (e.g Germacranolide), then by their subclass (e.g.
Heliangolide) and finally by their substitutional features. In Herout’s model, the concentric rings indicate the number of stages (and hence enzyme systems) the compounds must pass through in their formation. This of course presumes that
40 speed of formation of new biosynthetic enzyme systems are relatively constant, which might not be the case. However, the above chart did seem to overlap well with those that were based on morphological characters (Herout, 1971).
BA
ARNICA SENECIONEAE AM ER GU XA AM CICHORIEAE SA PS AMBROSIEAE GU
SA GU GE CARDUEAE SA GE HELIANTHEAE GE HYPOTHETICAL VE AM GU GU XA PROTOTYPES GE AM SA EUPATORIEAE GE GE VERNONIEAE INULEAE GE GE AM ANTHEMIDEAE VE XA EL GU SA
Figure 1.18 – Evolutional relations of tribes of Asteraceae
As proposed by Herout (1971)
GE – germacranolides EL – elemanolides SA – santanolides ER – eremophilanolides BA – bakkenolides GU – guaianolides AM – ambrosanolides XA – xanthanolides PS – psilostachyanolides VE – vermeeranolides
Although prior to Herout’s model, Arnica was initially placed in the tribe
Senecioneae, a number of physical differences (including opposite leaves and the receptacle with bristles) combined with the identification in Arnica of a number of sesquiterpene lactones including pseudoguaianolides (Rybalko et al. 1965,
41 Poplawski et al., 1971, Willuhn & Leven, 1995), not typical of the Senecioneae, led to it being removed from this tribe (Mabry and Bohlmann, 1977). Conversely, although some sesquiterpene lactones are found in several plant species and
Arnica produces pseudoguaianolides that are essentially the same as some of those in the genus Helenium, it should not be assumed that they are closely related.
A study into the chemistry and systematics of the genus Arnica by Willuhn et al.
(1995) which evaluated all the sesquiterpene lactones identified in Arnica species, found close affiliation between Arnica and the Gaillardiinae and Chaenactidinae
(within the Heliantheae), and suggested it be placed between these two sub tribes.
Although the Chaenactidinae have not so far been found to contain pseudoguaianolides, they do – like Arnica – contain phytomelan in their achene walls (Robinson, 1981). However, Willuhn et al. (1995) stressed that their work was designed to stimulate further research and that more analysis needed to be completed before Arnica could be placed with any certainty. The most up to date cladistic analysis by the Angiosperm Phylogeny Group has resulted in the following classification: Eudicots, Asterids, Euasterids, Asterales, Asteraceae
(APG, 2002) but as more species are subject to chemical profiling, it is possible that this position will be subject to change.
1.8.3.3 Sesquiterpene lactones in Arnica
The sesquiterpene lactones found in the Arnica species are extremely diverse which is rare for a single genus. Almost 70 different sesquiterpene lactones have been identified in Arnica (Willuhn & Leven, 1995), since the first was discovered in
42 the mid 1960s (Rybalko et al. 1965). Most of the sesquiterpene lactones in A. montana and A. chamissonis are of the pseudoguaianolide helenanolide type, the most common being helenalin, dihydrohelenalin and their derivatives (Table 1.3 and Fig.1.19).
The levels of total sesquiterpene lactones in both A. montana and A. chamissonis vary and have been reported at levels ranging from 0.31 % to 1.01 % w/w. There is a type of A. montana that contains mostly dihydrohelenalin esters and a central
European type that contains mainly helenalin esters (Willuhn et al.1994). Such plants that are identical morphologically, but which have dramatic differences in chemical constituents, are known as chemotypes (or chemovars) (e.g. Abraham et al. 1968).
Recent research on Spanish chemotypes have revealed that samples from heathland at high altitude have a similar helenalin content to the central European type, whilst those from peat bogs and meadows have higher levels of dihydrohelenalin (Perry et al. 2009). It has been suggested that for standardisation purposes, the extract for herbal preparations should not contain less than 0.4% w/w of total sesquiterpene lactones calculated as tigloyl helenalin with reference to the dried drug (Willuhn, 1998) quantified by an approved chromatographic procedure with santonin as an internal standard (Pharmacopoeia, European
2002).
43
Table 1.3 – The ester groups of Arnica’s sesquiterpene lactones
Structure Type Abbreviation Mass O
Acetate Ac 43 CH3
O
CH3 Isobutyrate Ibu 71
CH3 O
CH2 Methacrylate Meac 69
CH3
O CH3 Isovalerate Ival 85
CH3
O CH3 Senecioate Sen 83
CH3 O
2-methylbutanoate 2Mebu 85 CH3
CH3 O
Tiglate Tig 83 CH3
CH3 O CH3
Angelate Ang 83
CH3
44
1 Helenalin a- R = H (Helenalin) b- R = Acetate (Ac) c- R = Isobutyrate (Ibu) d- R = Methacyrlate (Meac) e- R = 2-methyl butanoate (2Mebu) f- R = Isovalerate (Ival) g- R = Tiglate (Tig) h- R = Angelate (Ang) (angelic acid ester)
2 Dihydrohelenalin a- R =H (11 α 13-dihydrohelenalin or plenolin) b- R =Ac c- R =Propanoate (Prop) d- R =Ibu e- R =Meac f- R =2Mebu g- R =Ival h- R =Tig i- R =Ang
3 Arnifolin a- R =Tig b- R =Ang c- R =Senecioate (Sen)
4 Dihydroarnifolin a- R =Tig b- R =Ang c- R =Sen d- R =Ival
5 Chamissonolide a- R =H R’ =H R’’ =OH b- R =Ac R’ =H R’’ =OH c- R =Ac R’ =H R’’ =OAc d- R =Ac R’ =Ac R’’ =OAc e- R =Ac R’ =Ac R’’ =H f- R =Ac R’ =H R’’ =H g- R =H R’ =H R’’ =H
Figure 1.19 - Diagram of helenalin, dihydrohelenalin and their esters
Willuhn & Leven, 1991
45
Douglas et al. (2004) found that different parts of A. montana flowers contained varying levels of sesquiterpene lactones. They found that acetyl dihydrohelenalin levels were highest in the stems of Arnica plants although the stems contain the lowest level overall of the sesquiterpene lactones. Schmidt et al. (1998) found that in very young plants helenalin derivatives dominate, but as the leaves develop, the levels of these compounds decrease as the levels of dihydrohelenalin type compounds increase. In the flowers, tigloyl helenalin and methacryl helenalin were the most abundant, followed by acetyl helenalin, isobutyryl helenalin, 2- methylbutyryl helenalin, isovaleroyl helenalin and then acetyl dihydrohelenalin.
Poplawski et al. (1971), found dihydrohelenalin, tetrahydrohelenalin and arconolides A,B,C,D & E in the leaf extract of A. montana and Zakharov et al.
(1971) identified arnifolin in the floral extract of both A. montana and A. foliosa.
Malarz et al. (1993) and Schmidt et al. (1998) both demonstrated that the accumulation of sesquiterpene lactones was associated with the differentiation of above ground parts, as there were none present in the root cultures. It is believed that their production is restricted to aerial parts, where they are deposited in glandular trichomes on the epidermal surface (Willuhn et al. 1986).
Schmidt et al. (1998) also found that while young shoots accumulated mainly helenalin derivatives, this decreased to near zero at the onset of leaf formation (3-
4 weeks). The dihydrohelenalin compounds increased at the same rate and remained constant for longer. The authors suggest that this is due to the presence of a hydrogenase system that converts the sesquiterpene lactones from the hydro
46 to the dihydro type and suggest that further investigation into this pathway would be of interest.
Willuhn & Leven (1991) found the total sesquiterpene lactone content for Arnica montana floral extract to be 0.85% (s[%]=3.2, n=8) and from Arnica chamissonis to be 0.87% (s[%]= 3.3, n=11) (excluding the pharmacologically non-equivalent chamissonilides). In flowers harvested in New Zealand, the total content of sesquiterpene lactones have been shown to increase with development from buds until the petals withered (Douglas et al. 2004). If this is shown to also be the case for plants grown in the northern hemisphere, as is likely, then it may be the case that harvesting should be left later than the currently recommended ‘partially open’ stage.
Douglas et al. (2004) also determined that one simulated mechanical harvest yielded twice as much plant material as two sessions of hand harvesting, although just over a third of this was stem material. When analysed it was determined that the stem material reduced the overall level of sesquiterpene lactones. However, when these were removed by rubbing (with a fixed 5.2mm screen to separate the stems from the flower parts), there was no significant difference in the levels of sesquiterpene lactones between the hand harvested flowers and those from the simulated mechanical harvesting treatment. This is likely to be due to the sesquiterpene lactone levels increasing with the age of the flowers. A mechanical harvest at ‘key times’ would include these, whereas hand harvesters would not, as only partially open flowers would be harvested.
47
1.8.3.4 The role of sesquiterpene lactones in plants
Sesquiterpene lactones can cause enzyme deactivation (e.g sulphydryl phosphofructokinase, Hanson et al. 1970) and hence may have a role as antifeedants. NMR spectroscopic investigations of the structure of helenalin, plenolin and chamissonolide (Schmidt, 1997) have shown that although all of them form conjugates with glutathione (GSH) through a Michael addition process (Fig
1.20). Chamissonolide with just the one alkyl group, reacted with less GSH than helenalin or plenolin. GSH is required to prevent excessive accumulation of lipid peroxides and is particularly susceptible to nucleophilic attack due to the sulfhydral group of the cysteine residue.
Guillet et al. (2000) found that helenalin reduced levels of GSH in Manduca sexta .
The helenalin in this case was purified from Helenium amarum (Raf.) H. Rock, but it is likely that helenalin plays a similar role in Arnica . In their study helenalin, the only sesquiterpene lactone investigated which had an α-methylene-γ-lactone and cyclopentenone electrophilic centre, caused the greatest depletion in GSH. The greater activity of helenalin could be explained by its possession of two alkylant structures and by the high flexibility of the ring skeleton which should allow it to react with a larger number of target molecules in a broad variety of chemical environments. (Kupchan et al. 1971 and Schmidt, 1996).
48
O O O
O O O
H
S S S H H R R R
H H R H R S S S
O O O
Figure 1.20 - The Michael addition process
Adapted from Schmidt, 1997
1.8.3.5 The medical uses of sesquiterpene lactones
Arnica extract has been used for many years for a wide range of ailments, from bruises, to heart disease and fevers (Flückiger & Hanbury, 1874). It is believed that the active compounds within this extract are sesquiterpene lactones, which have been shown to have antimicrobial, antimycotic, analgesic, anti-inflammatory and antiarthritic properties among others (Raison et al. 2000).
Sesquiterpene lactones from both dried (Wagner & Merfort, 2007 and Siedle et al.
2004) and fresh (Kos et al. 2005) A. montana material have been investigated for their anti-inflammatory properties. Klaas et al. (2002) found that a tincture of mostly helenalin esters was more active in anti-inflammatory assays than a tincture of mostly dihydrohelenalin esters. But for an anti-inflammatory treatment to
49 be effective, it first has to cross the greatest obstacle to transdermal diffusion which is the skin. In a skin permeation study by Wagner et al. (2004a) the permeation of extracts of both the most common Spanish (mainly 11 α-13- dihydrohelenalin esters) and the ARBO chemotype (predominately helenalin isobutyrate) were examined. They found that helenalin isobutyrate did not penetrate as far as the dihydrohelenalin acetate when applied ‘pure’ in ethanol/water. However, when applied in tincture form, permeation was enhanced and no clear difference could be detected between the Spanish and Arbo chemotype.
El Katten et al. (2001) showed that thymol and other monoterpenes can cause an enhancement effect, although it could also be that essential oil fatty acids also contribute (Wagner et al. 2004a). In a more recent study, Wagner and Merfort
(2007), the permeation of a gel and two ointment preparations was investigated, using a stripping method with adhesive tape and pig skin as a model. They found that the level of sesquiterpene lactone permeation through the stratum corneum was similar for all three preparations, although the permeation of the gel based sesquiterpene lactones decreased after 4 hours whereas that of the ointment remained fairly constant. They also found that the permeation of the sesquiterpene lactones was dependent upon formulation and total sesquiterpene lactone content, not the composition of the sesquiterpene lactones in the preparation.
Ly β et al. (1998) and Raison et al. (2000) found that helenalin, 11 α-13- dihydrohelenalin and their derivative esters inhibited the activation of transcription
NF-κB which is a mediator in the process of inflammation and a regulator of cell
50 death (Gertsch et al. 2003). Klaas et al. (2002) and Siedle et al. (2004) demonstrated that the inhibitory effect is dependent upon the ester groups of the sesquiterpene lactones. Those with unsaturated acyl moieties such as methacrylate and tiglinate have been found to display greater activity than the acetate derivatives.
Inhibition of NF-κB transcription has a positive effect on the occurrence of inflammation, particularly in the case of rheumatoid arthritis. Helenalin is bifunctional in that it has been found to both inhibit pro-inflammatory gene expression and also down regulated mRNA in a study of Jurkat T-cells and human peripheral blood cells. In the same study, acetyl dihydrohelenalin was found to moderately induce apoptosis in Jurkat T-cells whilst chamissonolide has been found to have cytoprotective effects (Gertsch et al. 2003).
In a study by Wagner et al. (2004b) three different sesquiterpene lactones (acetyl dihydrohelenalin, methacryl dihydrohelenalin and isobutyryl helenalin) were combined with human blood proteins in three different matrices (human serum albumin, plasma and whole blood). It was found that the extent of protein binding varied by sesquiterpene lactone in the order methacryl dihydrohelenalin < acetyl dihydrohelenalin < isobutyryl helenalin although the interactions of sesquiterpene lactones with plasma proteins is not fully understood.
Sesquiterpene lactones are known to bind covalently to sulfhydral groups of biological molecules by Michael addition of their α, β unsaturated carbonyl structures (Schmidt, 1996) which can cause enzyme deactivation (Lee et al.
51 1977b). Thiol groups of enzymes are particularly prone to this reaction and sesquiterpene lactones have been shown to inhibit sulfhydral group bearing enzymes such as phosphofructokinase (Hanson et al. 1970).
The antitumor effects of helenalin were first investigated in the 1970s when
Kupchan et.al (1970) demonstrated that the O=C-C=CH2 of the α-methylene-γ- lactone system can act as an alkylating centre in cytotoxic anti-tumour lactones.
However, later work by Geran et al. (1972) demonstrated that compounds such as tenulin (Fig 1.21) and cyclopentanone, which do not have a α-methylene-γ-lactone moiety, were effective anti-tumour agents too, suggesting that the cyclopentenone ring is also important. Waddell et al. (1983) found a strong and exclusive Michael- like selectivity of this ring for sulphur nucleophiles. Lee et al. (1977a) suggested that helenalin inhibited protein synthesis in tumours because they underwent a
Michael addition with the sulphydryl groups of reduced glutathione and L-cysteine producing a structure like that in figure 1.22. The term ‘bifunctional’ is now applied to sesquiterpene lactones with two alkylating centres (Recio et al. 2000).
However, lipophilicity is also important as this property can mean the sesquiterpene lactone will penetrate the membrane more easily and hence would increase cytotoxicity. For helenalins there is a size limit on such groups beyond which the toxicity of the sesquiterpene lactone decreased. In contrast, for dihydrohelenalins, cytotoxicity was directly proportional to the size of the ester side chain, likely due to its distance from the alkylating centre (Beekman et al. 1997).
52 H
O
O O O
HO
Figure 1.21 – Tenulin
Waddell, 1982
R H R = S-CH2-CH-NH-CO-CH2CH2-CH-COOH | | C=O NH2 O | NH | O O CH2 | COOH
R
Figure 1.22 - The product of reacting glutathione with the α-methylene group of the
γ-lactone and the double bond of the cyclopentenone ring
As proposed by Lee et al. 1977b
The potential of helenalin to act as an inducer of HL-60 differentiation has also been investigated (Kim et al. 2005). Cancer cells generally lack a capacity to develop into mature non-replicating adult cells, hence they proliferate. If the cancer cells are triggered into differentiation, then they will stop replicating and eventually undergo programmed cell death. Kim et al. (2005) found that helenalin induced differentiation of HL-60 leukaemia cells into granulocytes, although the exact mechanism of action was not clear. In addition, Huang et al. (2004) demonstrated
53 that helenalin inhibited telomerase in vitro , likely reacting with the cysteine residue of the telomerase proteins. This has the potential to explain the anti-tumour activity of helenalin. However, Woerdenbag et al. (1995) found that flavonoids from the
Arnica decreased helenalin-induced cytotoxicity in a human lung carcinoma cell line.
In an investigation of guinea pig heart tissue, Takeya et al. (1983) examined the effect of helenalin on strips from the left atrium and papillary muscles from the left ventricle. The contractility of both improved irreversibly (did not return to control after washing) which demonstrated the potential of Arnica extract as a cardiotonic
(Robles et al. 1995).
Dihydrohelenalin (plenolin) and arconolide (both often found in A. chamissonis ) were demonstrated to have antimicrobial activity when extracted from Centipeda minima (L) A.Br and Aschers (Asteraceae), a plant used to treat sinus infections in
Nepal (Taylor & Towers, 1998).
In its homeopathic form, Arnica has been demonstrated to be effective at reducing bruising if administered to rats before the application treatment of an oedema causing agent (carraggeenin) (Macêdo et al. 2004). Lussignoli et al. (1999) also found that a similar preparation containing A. montana extract ‘speeded up’ the healing process in rat paw oedemas.
54 1.8.3.6 Chemical synthesis of sesquiterpene lactones
To date, there have been no reports of a cost effective route for the chemical synthesis of the sesquiterpene lactones present in Arnica . Money & Wong (1996) have identified an enantiospecifc method for the synthesis of pseudoguaianolides.
The process, involving the synthesis of hydroazulenoid ketones from camphor, represented a method for the production of sesquiterpene lactones such as mexicanin, bigelovin and helenalin. Such methods of production could be useful when producing cytoxins as flavonoid contaminants have been shown to reduce cytotoxity of helenalin (Woerdenbag et al. 1995). However, this method is expensive and time consuming, and as mentioned previously, studies by Wagner et al. (2004a) indicate that compounds found in the tincture of Arnica can act as penetration enhancers for the 'active' sesquiterpene lactones when used as an anti-inflammatory. It may be that other compounds in the Arnica extract are boosting its effectiveness and hence the plant extract could be more effective than the pure compounds alone (Wagner et al. 2004a). Indeed, using higher concentrations of the active compound alone to compensate may induce irritation and allergic responses.
1.8.4 Other constituents of Arnica extract
Arnica has been reported to contain a wide range of compounds including: flavones, chromenes, diterpenes, pyrrolizidine alkaloids, inositol esters (Willuhn &
Leven, 1991), astragalin, betaine, choline, cynaroside, inulin, luteine, phytosterol, scolimoside, scopoletin, trimethylamine, umbelliferene, flavonoid aglycones, flavonoid glycosides, caffeoylquinic acids and xanthophylls (Duke et al. 1985,
Merfort 1988 and Merfort 1992). Arnica root has been reported to contain lignans
55 (Willuhn & Leven, 1991) arnicin, arnidiol, anthoxanthine and tannin along with a yellow-green fluorescent flavonic pigment that is also found in the flowers (Duke et al. 1985). The leaves have been reported to contain dihydrohelenalin and its esters (arnicolides) (Poplawski et al., 1971).
An essential oil can be extracted (0.2-0.35%) which is reported to contain ca. 40-
50% fatty acids and 9% alkanes (C 19 –C30 ) (Kating et.al, 1970), thymol derivatives and mono and sesquiterpenes (including α-phellandrene, myrcene, humulene, δ- cadinene, caryophyllene oxide, (Lee et al. 1977a & Willuhn & Leven, 1991) cinnamic acid and derivatives (including chlorogenic acid and caffeic acid), coumarins (umbelliferone and scopoletin), polyacetylenes, choline and xanthophylls (Raison et al. 2000). The essential oil has been reported to have antiseptic activity (Schroder & Merfort, 1991)
Apart from sesquiterpene lactones, the flavonoids are the most thoroughly investigated secondary metabolites in Arnica (Willuhn & Leven, 1991). There are three major groups, the flavanones, flavones and flavonols. The flavonoids of
Arnica are generally mono or dioxygenated in ring B, and in the glycosides the sugar is normally at the 3-position and the flavones at the 7-position. However, in
Arnica chamissonis there are acylated glycosides, where the flavonol glucosides are esterified with acetic acid and the flavone glucosides with 2-methylbutyric acid
(Willuhn & Leven, 1991).
Lignans have been found in the flower heads of Arnica flowers (specifically,
A.angustifolia ssp attenuata (Greene) Maguire) and in the roots of A. montana and
56 A. chamissonis (Spitaler et al. 2006) . In particular a rich variety of phenyl propanoids were found in the roots of A. chamissonis in significant amounts (Ibid).
Flavonoids present in Arnica extract include isoquercitrene, luteolin, luteolin-7- glucosides and astragalene (Raison et al. 2000 & Roki et al. 2001).
1.8.5 Toxicology
Extracts of A. montana have been shown to be non-toxic in acute toxicity tests on rabbits, mice and rats (Anon, 2001). However, it was shown to be mutagenic in an
Ames test, (although no carcinogenicity, reproductive or developmental toxicity data is available) and the Food and Drug Administration (FDA) classifies A. montana as an unsafe herb, although it can be used as a flavour in an alcoholic preparation (FDA, 2005). It notes that both the aqueous and alcoholic extracts contain choline and "two unidentified substances which can produce violent toxic gastroenteritis, nervous disturbances, change in pulse rate, intense muscular weakness, collapse and death" (FDA, 2005). Macêdo et al. (2004) reports that when A. montana is ingested it can cause vascular dilation, blood stasis and increased capillary permeability. By disrupting the endothelia membrane it can induce haemorrhaging and in muscle it can induce pain similar to ‘stitch’ pain, induced by excessive exercise.
Despite the lack of sensitisation on animal skins, those who work regularly with
Arnica montana have reported a delayed type IV allergy that is referred to as
Arnica dermatitis (Anon, 2001 and Spettoli et al. 1998) and it has elicited positive photopatch test results (Victor et al. 2010). Both are believed to be due to
57 helenalin (Duke et al. 1985), and other sesquiterpene lactones (Hausen, 1996 and
Spettoli et al. 1998), although Raison et al. (2000), also state that it is essential that the proportion of lipophilic flavonoids algeics in the flowers is low and that when used externally glycosides do not penetrate the skin. In contrast, recent work by Lass et al. (2008) found tinctures and sesquiterpene lactones from both
Spanish and central European Arnica were unable to induce hypersensitivity in mice with intact immune systems, attributed to the anti-inflammatory effect of the extracts. They recommended that sesquiterpene lactones and Arnica tinctures be classified as weak contact allergens.
The recommended dosages as prescribed by the Kommission E in their monograph for Arnica e Flos (1984) can be found in Table 1.4.
Table 1.4 – A. montana dosage (Kommission E, 1984)
Mixture 2.0g drug in 100ml water Tincture (for dressings) 3-10 fold dilution in water Tincture (for mouthwashes) 10 fold dilution in water For ointments Max 20-25 tincture Arnica oil I part drug to 5 parts saturated vegetable oil Ointments Max 15 percent Arnica oil Contraindications None known
Whilst many sesquiterpene lactones have cytotoxic properties, many of them also have unwanted toxicity to normal cells, allergenic properties or have poor bioavailability due to their interactions with plasma proteins.
58 1.9 Discussion
As outlined in this chapter, considerable work remains to be completed in order to fully elucidate the biosynthetic pathways of the sesquiterpene lactones in Arnica and the levels of natural variation in their expression. It is possible in light of the endangered status of A. montana in the wild, that genetic engineering methods
(such as the over expression of a chimeric farnesyl diphosphate synthase gene which boosted sesquiterpene lactone production in Artemisia annua (Chen et al.
2000)), will eventually be applied to this species. While it is unlikely that such a modified plant would be commercially viable as a crop, given that the market for the herbal extract is traditionally organic with a non-interventionist approach to nature, the unexpected consequences of such a modification, such as those observed in Mint (Mahmoud & Croteau, 2001), could reveal previously unknown intermediates and pathways and provide valuable biosynthetic insights. This would not only allow better understanding of plant defence pathways, but would also provide chemosystematic insights, such as the evolution of the Asteraceae in relation to its sister families.
The primary focus of biomedical companies is at the moment on the production of pharmaceuticals via a microbial vector. This is due to less flavonoid
‘contamination’ and such vectors being quicker and easier to screen than whole plants (Harborne, 2000). However, phytochemical research in a medicinal context is increasing (Borris, 1996) and it may be that modified versions of Arnica with boosted sesquiterpene lactone production will be available in the future.
59 Till then, it is possible that investigations into the causes of natural levels of variation in active compound levels could provide clues as to how to boost their production. This is also important if a stable, commercially viable crop yield is to be maintained. This study investigated the effect of species and a number of agronomic treatments on the yield and chemical composition of the Arnica cultivated in Orkney . Sesquiterpene lactones have been determined to be the most relevant compounds from the pharmaceutical, ecological and the taxonomic context for this species and were hence the main focus of this research. If these compounds have the role of defence or stress compounds in Arnica as they are in other species (eg. Guillet et al. 2000 & Saker et al. 2000), it may be their levels could be increased by altering the plant’s environment via changes to the agronomic treatment. For such conditions to be identified, field trials are essential.
Whilst there have been many trials investigating the extracts from Arnica (e.g
Kating et al. 1970, Leven & Willuhn 1987 and Merfort, 1992 ), their medicinal properties (e.g. Klaas et al. 2002, Kim et al. 2005, Alfredo et al. 2009), ideal agronomic regimes (e.g Bezzi & Ghidini, 1989, Bomme & Daniel, 1994 and
Buthuc-Keul & Deliu, 2001) and their ecological context (e.g. Bryant et al. 1983,
Bruelheide & Scheidel, 1999 and Kahmen & Poschlod 2000) there have been limited studies to date that investigate more than one of these factors and how they interrelate (Douglas et al. 2004, Spitaler et al. 2006 and Smallfield and
Douglas, 2008). This programme aims to contribute to the literature by investigating both chemical and agronomic factors that influence the productivity of both A. montana and A. chamissonis .
60 1.10 Research question
How do varied provenance, environmental and agronomic factors influence the quantity and range of active compounds found in Arnica montana and Arnica chamissonis ?
1.11 Aim
The literature reviewed in this study indicates a clear need for a better understanding of the growth and quality attributes of cultivated Arnica. Hence the aim of this research is to investigate the environmental and agronomic factors that influence the yield of A. montana and A. chamissonis and to determine the nature of any relationships between these factors that may indicate the reason for their production.
1.12 Objectives for this study
There were a number of objectives to this study which was a reflection of the scope and breadth of the research question. They were:
• To compare A. chamissonis extract with A. montana in terms of oil yield and
active compounds.
• To determine the effect of agronomic conditions on crop yield and quality
• To provide an insight into the role of the active compounds within the plant.
• To determine the ‘ideal’ regime for Arnica grown in Orkney.
• To apply qualitative and quantitative analyses of the flower material grown
in Orkney to determine the quality of the product and utilise yield to judge
the feasibility of the next stage of production.
61 • To compare the quality of Orkney grown material with that available on the
herbal market and assess whether Orkney grown Arnica has any potential
as a commercial high value extract crop .
62 2 The extraction, identification and analysis of sesquiterpene lactones
2.1 Introduction
Arnica flowers have been reported to contain, among other compounds, flavones, diterpenes, sesquiterpenes and essential oil consisting of fatty acids, paraffins, phenols, esters, alkanes (Kating et.al, 1970), monoterpenes and sesquiterpenes
(Kating et al. 1970, Willuhn, 1972 (a,b), Willuhn & Leven, 1991). The aim of this study however, was to focus on the levels of sesquiterpene lactones in Orkney produced Arnica and to determine if their levels were affected by variations in agronomic regime. Consequently, an analytical methodology was required for the
Arnica material grown in the Orkney field sites. This chapter describes the factors influencing the final extraction method, the methodology of separation and the process by which compounds were identified and quantified.
2.1.1 Flower material
Although fresh flowers are used in the preparation of homeopathic remedies (Kos et al. 2005), in the preparation of Arnica tincture, the use of dried flowers is more common (e.g. DAB 7, 1964 and DAB 10, 1991). In addition all flower material in this study would be harvested in Orkney for later analysis in Inverness and hence a method which allowed for storage of the flowers was required.
As the proposal involved the analysis of a considerable amount of flower material, the method of extract production and analysis had to allow for quick turn around, yet not lead to the damage of material, or the loss of storage stability. This is not always the case in method development, where often the emphasis is on
63 producing high quality extract to allow the identification of trace components.
Method development stages which take high throughput into account are hence also important in any study where the aim of the research is to quantify known components, rather than identify new compounds.
It was also vital for a comparison of the sesquiterpene lactone content in the
Orkney grown material to be made with that of both commercially available material from Europe and that reported in the literature. This would then make it possible to judge the quality of the Orkney grown material.
The idea floral stage of material harvested for analysis varies between partially open flowers (e.g DAB 10, 1991) and over flowers (Douglas et al. 2004). It was hence important that flowers at different stages be harvested to allow for the ideal time of harvesting to be identified.
2.1.2 Extraction
Sesquiterpene lactones in Arnica extract are recommended to be within the range of 50-170 µg/ml (0.50-1.7% w/w) (Willuhn & Leven, 1995) or at least 0.40% w/w calculated as helenalintiglinate for topical Arnica preparations (Pharmacopoeia,
European 2000 and Willuhn & Leven, 1995). Such topical preparations contain the tincture of Arnica blossoms but this tincture can vary in content from 5-25% v/v
(Bilia et al. 2006). It is important when producing extracts that the content of sesquiterpene lactones is quantified so that the correct amount of tincture can be added to the ointment/gel to ensure efficacy.
64 Although a number of methods for the extraction and analysis of sesquiterpene lactones from Arnica extract have been developed (Leven & Willuhn, 1987,
Willuhn & Leven, 1991, Douglas et al. 2004, and Kos et al. 2005), none have been required to process the number of samples proposed in this study. When such a large number of repetitions are required, repeatability, reliability of results and costs of analysis inevitably have to be considered. To this end, several methods were considered for the solvent extraction of Arnica flowers.
A particularly useful starting point for analytical methods are pharmacopoeias.
Published by the authority of a medical or pharmaceutical group, pharmacopoeias contain guidelines for the identification and preparation of medicines
(Braunschweig, 2008). They contain regulations and procedures from the different fields of pharmacy, in addition to monographs which define the testing methods for each drug, as well as instructions on how to identify, purify and store them. Arnica is listed in the pharmacopoeias of at least 13 countries, although the German pharmacopoeias are the most commonly referred to in the literature (e.g Schröder
& Merfort 1991, Willuhn & Leven 1995 and Wagner & Merfort 2007). The German
Pharmacopoeia ( Deutsches Arzneibuch, or DAB ) is also known as the German dispensatory and has so far appeared in ten volumes (Table 2.1).
65 Table 2.1 - The ten editions of the German Pharmacopoeia
Braunschweig, 2008
Title DAB Year of publication Pharmacopoea Germanica, edition I DAB1 1872 Pharmacopoea Germanica, editio altera DAB2 1883 Pharmacopoea Germanica, editio III DAB3 1890 German dispensatory 4 DAB4 1900 German dispensatory 5 DAB5 1910 German dispensatory 6 DAB6 1926 German dispensatory 7 DAB7 1964 German dispensatory 8 DAB8 1978 German dispensatory 9 DAB9 1986 German dispensatory 10 DAB10 1991
Whilst DAB 6 permitted the preparation of medicinal Arnica tincture by maceration, from DAB7 onwards the recommended method was percolation (DAB 7, 1964).
This required a large number of whole flower heads and led to a slow extraction rate (Wichtl, 1987). After numerous complaints about this method, an alternative was developed (DAB 10) which permitted the preparation of tincture from shredded or roughly powdered flower heads (DAB 10, 1991).
In 1987 Leven & Willuhn described a method (DAB 9) for the extraction of neutral and amphiphilic sesquiterpene lactones from Arnica chamissonis . In some pharmacopeias, such as those from GDR and Russia, there is no distinction between the use of A. montana and A. chamissonis. Despite this seeming equivalence, It was not until DAB 9 that the use of A. chamissonis was permitted in the German pharmacopoeias and even then only A. chamissonis subsp. foliosa
(Schröder & Merfort 1991).
66
In the DAB 9 method for analysis of this material, 1.000g of dried flowers was extracted twice in 50 ml of MeOH/H 2O (1:1 v/v) using a reflux condenser (50°C) with stirring. Reflux extraction is both time and energy expensive and in this case appeared to increase the levels of standard deviation. The concentrated Arnica extract was then applied to an Extrelut™ column and eluted with 100 ml
DCM/EtOAc (1:1 v/v) to remove hydrophilic compounds.
Although the use of the Extrelut™ column was no doubt effective, this would add another potential source of error. The Extrelut™ extracts in the Leven & Willuhn
(1987) method, were evaporated to dryness, resuspended in 10 ml MeOH and diluted with 10 ml H 2O. This was then shaken with 7g Al 2O3, centrifuged and filtered. 10ml of the solution was taken, evaporated to dryness, dissolved in 2ml
MeOH:H 2O, filtered via Sartorius filters, then 0.5ml evaporated and dissolved in
0.25 ml of EtOAc. The latter part of the method was designed to produce samples that could be run both on HPLC and GCMS but could be considered excessive for the purposes of sample comparison as well as a source of potential error. The sesquiterpene lactone concentration in ethanol tincture has been shown to decrease by 37% at 30°C, 32% at 25°C and 13% at 4°C after three years storage
(Schmidt et al. 2000) so storing samples at 4°C or lower is important.
Willuhn & Leven (1995) analysed the differences in the sesquiterpene lactone content of A. chamissonis extract produced by the DAB 10 method, by maceration and by infusion. Although they describe a 16% increase in the proportion of sesquiterpene lactones passing into the water-alcohol solution via percolation, the
67 standard deviation was lower in the maceration method, and the percentage content of the individual components was similar for both (table 2.2). In addition, initial method development work indicated that the time and energy costs involved in extracting large numbers of samples via the percolation method would be prohibitive, reducing the number of samples it would be possible to analyse and reducing the scope of this study.
There is some variation in the amount of sesquiterpene lactones as reported in the literature, with levels of sesquiterpene lactones extracted from the blossoms of A. montana ranging from 0.31% (Willuhn et.al, 1994) to 1.04% (Spitaler et al. 2006), and 0.94% reported in withered flowers to 0.51% in buds (Douglas et al. 2004).
The Spanish chemotype of A.montana contains mostly dihydrohelenalin esters
(Willuhn & Leven, 1991), but work by Perry et al. (2009) indicates that this may be a consequence of habitat, with a significant positive relationship found between type of site and the content of sesquiterpene lactones. A. montana found on heathland sites at high altitude had significantly higher levels of helenalin esters, whilst those on meadows and peat bogs were mostly of the dihydro type.
Commercially available Arnica extract has been found to contain 0.64 mg/ml of sesquiterpene lactones consisting of almost 50% 11 α13-dihydrohelenlin methacrylate, corresponding to the predominant Spanish chemotype (Willuhn &
Leven, 1995).
In A. chamissonis flowers, sesquiterpene lactone content has been reported at levels ranging from 0.2% (Willuhn et al. 1994) to 1.7% (Leven & Willuhn 1987).
However, it seems that just as the most common Spanish A. montana contains
68 predominately dihydrohelenalin esters, there is a Spanish chemotype of A. chamissonis which has similar characteristics (Willuhn et al. 1994). Although to date no quantitative variation has been found in the sesquiterpene lactone content of flowers grown in central Europe, in a study of north American A. chamissonis, five chemotypes were identified from different locations on the west coast (Figure
2.1 and table 2.3) (Willuhn et al. 1994).
It would seem from the literature that the levels of sesquiterpene lactones generally are more variable in the extracts of A. chamissonis , although some variation (seemingly linked to the country of origin), has also been reported in A. montana (Willuhn & Leven, 1991). In a later work, Willuhn & Leven (1995), described the average sesquiterpene lactone content described for Arnica blossoms as 0.55% and from a flower weight of 2-3g, they estimated that a sesquiterpene lactone content of 0.01-0.016% could be obtained.
69 Table 2.2 - Percentage content of the sesquiterpene lactones extract of A. chamissonis flowers.
Adapted from Willuhn and Leven (1995)
Tincture Tincture ‘Standard Infusion Compound (DAB 10 – (Maceration) average’ percolation)
Dihydrohelenalin 4.5 10.4 9.8 12.1 Helenalin 14.0 7.8 7.2 7.3 Acetyl 2.1 0.7 1.0 2.1 dihydrohelenalin Chamissonolide 3.7 3.5 2.6 7.7 Dihydroarnifolin 12.8 8.9 8.9 9.7 Acetyl 12.5 10.8 11.0 11.4 chamissonolide Arnifolin 11.1 11.4 10.2 11.2 Arnifolin B 3.4 3.8 3.5 4.1 11,13- 0.2 1.8 0.9 0.9 dehydroflexuosin B 6-desoxy-4-O-acetyl 3.3 7.9 6.6 8.7 chamissonolide Tigloyl 18.7 14.3 15.5 11.2 dihydrohelenalin Angelicoyl 2.1 3.1 4.1 3.1 dihydrohelenalin Tigloyl helenalin 9.9 13.7 15.9 9.1 Angelicoylhelenalin 1.7 1.9 2.8 1.4 100 100 100 100 Helenalins 25.6 23.7 25.9 17.8 Dihydrohelenalins 27.4 28.5 30.4 28.5 Arnifolins 14.7 17.0 14.6 16.2 Dihydroarnifolins 12.8 8.9 8.9 9.7 Chamissonolides 19.5 22.2 20.2 27.8
70 Table 2.3 - The five chemotypes of A. chamissonis as identified by Willuhn et al. (1994), and their source locations
Chemotype of A. Characteristic content Location chamissonis Mono Lake, Quartz Path and A Mostly dihydrohelenalins Grass Lake Mostly dihydrohelenalins and Lake Tahoe and Trukee B helenalins River Mostly dihydrohelenalins and Fallen Leaf Lake and Convay C dihydroarnifolins Summet D Mostly arnifolins Crazy Lake (CL) E St Mary Lake and Carson Mostly chamissonolides Pass
In this study, due to the sometimes limited amount of material available, and the possible need for repeat extractions, a flower weight of just 1g was available for analysis. As any methodology would need at least to be able to detect levels of the active compounds in the sample from 0.005-0.008% or 0.08 mg/ml, the method development stage of the investigation took this into account (section 2.2.2.2).
With high levels of natural qualitative and quantitative variation, small amounts of sample material and large number of replicates, it was vital in this study that the reliability of the results be ensured. Good Laboratory Practice (GLP) is a system by which laboratory studies can be planned, performed and completed while ensuring that the data are a true reflection of the results obtained (GLPMA, 1999).
To this end, GLP was ensured wherever possible and where it was not, compensatory methods were employed in the method development and are explained in the methodology.
71
Figure 2.1 - Sample sites of North American A. chamissonis in the study by Willuhn et al. 1994
Google Maps and Google Earth, accessed 29/01/09
Both A. montana and A. chamissonis produce very little essential oil and hence few studies have investigated it further. Solvent extraction of Arnica flowers requires less flower material, less time and less energy (in the form of heat) and so is favoured by the vast majority of analyses (e.g. Willuhn et al. 1994, Willuhn &
Leven 1995 & Douglas et al. 2004,). In addition, as less heat is required employing solvent extraction, less degradation of metabolites can be predicted. This study however attempted to investigate both the quality and the quantity of the essential oil extracted from the flowers and roots of the plants, as studies of other species have shown the content of essential oil to be altered by agronomic treatment (e.g.
72 Klimánková, 2008, Fuente et al. 2003 and Alvarez-Castellanos & Pascual-
Villalobos, 2003) and others have shown that components of essential oil (e.g. thymol and fatty acids) can aid permeation of the active compounds through the epidermis (Wagner et al. 2004a and El Katten et al. 2001).
In an attempt to ascertain whether there is a relationship between the agronomic treatment of Arnica and the essential oil produced, a methodology for oil extraction was developed. This oil was then used in the development of a qualitative and quantitative method of analysis for the essential oil produced from differing fertiliser and weeding treatments.
2.1.3 Chemical analysis
A number of methodologies for both the extraction of Arnica and for the analysis of the resulting material have been outlined in the literature. In part, the wide variation in analytical methodologies reflects the different methods of analysis available to different teams of researchers, but there are additional factors to be taken into consideration.
Whilst reverse phase high performance liquid chromatography (or high pressure liquid chromatography - HPLC), is a very flexible system, problems can arise with the UV absorption of some A. chamissonis sesquiterpene lactones (Leven &
Willuhn, 1987) which lack a chromophore (UV or visible light absorbing groups) and so cannot be analysed effectively by HPLC with UV detection at 254nm. In such cases, although short wavelength UV monitoring is an option, this in turn limits the number of solvents that can be used due to their own low level
73 absorption (Merfort, 2002). In addition, there can be difficulty in resolving methacryl helenalin and isobutyryl helenalin (Spitaler et al. 2006). Despite this, there exist many methods which utilise HPLC (e.g. Spitaler et al. 2006, Leven &
Willuhn, 1987, Willuhn & Leven, 1991, Willuhn et al. 1994, and Willuhn & Leven,
1995) and in a review of analytical techniques for sesquiterpene lactones Merfort
(2002) found that HPLC was considered the method of choice for the identification of sesquiterpene lactones due to their low volatility.
However, gas chromatography - mass spectrometry (GC-MS) has also been employed successfully (Schmidt et al. 1998, Malarz et al. 1993, Leven & Willuhn,
1987, Willuhn et al. 1994 and Willuhn & Leven, 1991), and most importantly for this study, the standard deviation has been found to be less when GC-MS is employed (Leven & Willuhn, 1987). LC-MS (or HPLC-MS) combines the separation power of HPLC with the analytical capability of mass spectrometry. It is a powerful technique with many applications, but unfortunately the costs associated with LC-MS were too high for this analysis. GC-MS facilities were readily available for the purposes of this study and so this, paired with a need for low deviation in results due to the quantitative nature of the analysis, and the high suitability of this method for the analysis of the volatile compounds of interest led to the selection of extraction methodologies outlined in section 2.2.2.
2.1.4 Compound identification
Although some work has been done on the composition of essential oil components of A. montana extract (Kating et al. 1970 & Willuhn, 1972a,b &c), more recent analysis by GC-MS has been limited. However, many of the key
74 components of essential oils are common to many species (e.g. limonene and α- pinene) (Adams, 2001) which makes it possible to identify the majority of essential oil compounds with reference to the work done on other oils. Adams (2001) has produced an extensive library of the mass spectra and retention times of essential oil compounds and so these were used in order to identify the main components of the oil.
In their GC analysis of solvent extracted sesquiterpene lactones, Leven and
Willuhn (1987) found that compounds with a 2-α-hydroxycyclopentanone structure were retained longer than analogous compounds with an α,ß-unsaturated ketone structure. As the GC system has higher resolving power than HPLC, the E,Z isomers should be well separated. However, it has been found (Leven & Willuhn,
1987), that the 11 α,13-dihydroarnifolin C peak can merge with the 6-O- acetylchamissonolide, 11,13-anhydroflexuosin B with Arnifolin, and 11 α-13- dihydroarnfolin with flexuosin B. This emphasises the importance of utilising mass fragments in confirming peak identity.
In their work on sesquiterpene lactones, Tsai et al. (1969) found that the characteristic mass fragment peaks for pseudoguaianolide sesquiterpene lactones were m/e 95, 96, 122, 123 and 124. They also found that all parent molecules showed weak peaks below their expected m/e values (e.g. m/e 244) which were characteristic of the loss of methyl groups, acetyl groups and of water. Indeed, when analysing germacranolides with hydroxyls and/or esters (like sesquiterpene lactones) by GC-MS, the parent ion can be missing altogether due to the loss of
75 either water or the side chain, via McLafferty rearrangements (Figure 2.2) (Fischer,
1978 and Merfort, 2002).
H R1 H R OH2 1 O +
R CH2 CH2 R
Figure 2.2 – McLafferty rearrangements
McLafferty, 1959
The high resolution mass spectrum of helenalin was first established by Tsai et al. in 1969. In their analysis, the accurate masses of the peaks of high m/e values unsurprisingly corresponded to the elimination of water (m/e 244) and carbon monoxide (m/e 234), subsequent methyl removal (m/e 229) and combinations of these. This was confirmed by later analysis and that helenalin and dihydrohelenalin esters could be identified by peaks of m/e 244 and 246 respectively (Wagner et al. 2004b and Wagner & Merfort, 2007). The most prominent peaks, were m/e 95, 96, 122, 123 and 124, which when examined at high resolution and compared with spectra of pseudoguaianolides devoid of the 4-,
6-, and 8-oxygenation pattern characteristic of helenalin provided supporting evidence for the cleavage pattern proposed in figure 2.3.
Tsai et al. (1969) proposed that helenalin ( 1) can form the intermediate a which can then fragment via the 9,10 bond to form b. Both helenalin and acetyl helenalin
(2) can form the intermediate c which can form the m/e 123 fragment d by cleavage of the 9,10 bond. Subsequent decarbonylation could then lead to the m/e
76 95 fragment k. The formation of the m/e 124 fragment was proposed to arise from proton transfer from c, via a six-membered cyclic transition to the oxygen stabilised e. Subsequent cleavage of 9,10 bond leads to the production of f.
However, they acknowledged that hydrogen transfer could also occur from other sites such as those outlined in figure 2.4. Tsai et al. (1969) also proposed that the m/e 151 fragment ( n), could arise via the 6,7 cleavage mechanism outlined in
Figure 2.5. Tsai et al. (1969) also proposed that the m/e 151 fragment ( n), could arise via the 6,7 cleavage mechanism outlined in Figure 2.5.
Such rearrangements can lead to the loss of the parent ion and so the use of the mass spectra in this case was limited to the identification of the basic skeleton and diagnostic peaks such as the ester side chains (Table 1.3).
Lactones which contain either hydroxyls or ester groups show weak or no parent ions as they can easily lose water or labile side chains (e.g. M-H2O or M-COOH).
Commonly the acylium ion of the ester group is the most intense peak (Banthorpe,
1991). An example acylium ion, in this case formed by the fragmentation of an acetyl ester group is shown in Fig 2.6.
77
1
# O * O m/e 122
b O O O O O HO O H a H
H O O
HO O O O O O c H ♦ * C CO H2 2 *
O H HO O O m/e 95 HO e l
# = cleavage of 5, 6
H ♦= cleavage of 1,10
HO HO * = cleavage of 9,10
d f
H HO m/e 123 HO m/e 124 + C H O + C7H11 4 8 k m/e 95 m m/e 96
Figure 2.3 – Proposed routes to ion fragments
Tsai et al. 1969
78
a)
m/e 124 (R=H) m/e 125 (R=D)
1 1
O H O
O O O O RO RO
R=H or D; Ac or COCD2
H
H H O O
m/e 123
b)
O O
O O RO O R O O
R2C CO
R=H or D Cleavage of 1,10
RO O
m/e 124 (R=H) m/e 125 (R=D)
Figure 2.4 – Formation of alternative m/e 123 (a) and124 (b) fragments through hydrogen transfer from alternative sites
Tsai et al. 1969
79 1
O H O
O O O O HO O HO HO n m/e 151
Figure 2.5 – Formation of fragment m/e 151 fragments
Tsai et al. 1959
O
CH3
Figure 2.6 – Acylium ion
2.1.5 Quantification
Ideally, the compounds to be analysed would have been isolated, purified and known concentrations of each would have been run alongside the extracts for comparison. Unfortunately, due to both the amount of material that would be required and a lack of commercially available standards, this method was not feasible for this study. Instead, published methodologies of analysis which were tested alongside small amounts of pure compounds were reviewed and the most appropriate adopted.
80 There are two main methods for the quantification of the constituents of the Arnica extract. Bilia et al. 2006 calculated their content using a calibration curve for santonin. However, in order to take into account the different MS-response factors of the sesquiterpene lactones, Leven & Willuhn (1987) adopted a different approach. The peak areas in the GC-MS were measured by multiplying the peak width at half height by the total peak height. The GC-MS calibration factors f(sl/st), were determined by taking into account the number of hydrogen bearing carbon atoms in the molecule and those in the standard compound, with the molecular weight of both.
Due to some compounds eluting together and the presence of an unidentified compound, the total amount of sesquiterpene lactones in their study was calculated by taking one component (normally the main compound) as a reference substance suitable for standardization requirements (Leven & Willuhn, 1987).
Determined by GC and calculated as helenalin, dihydrohelenalin (DH), 6-O-tiglyol
H, 6-O-tiglyol DH, arnifolin, 11 α,13-dihydroarnifolin and chamissonolide, the average total amount of sesquiterpene lactones was 7.20 mg/g with a relative standard deviation of 4.40% for GC.
Willuhn & Leven (1987) proposed that if the A. chamissonis material varied considerably in terms of compound composition (e.g. when sourced from different geographic regions), it would be more practicable to compare sesquiterpene lactones according to the substitution types, either by whether they are 11 α,13- dihydro or 11,13 anhydro types, or by the different types of substitution on the cyclopentane ring (table 2.4).
81 Table 2.4 - Partial amounts of the sesquiterpene lactones in the flower heads of A. chamissonis subsp. foliosa
Adapted from Leven and Willuhn, 1987
Compound Calculated as mg/g dry weight Helenalin Helenalin 1.03 Dihydrohelenalin (DH) Dihydrohelenalin (DH) 0.40 H-esters 6-O-Tiglyol H 0.46 DH-esters 6-O-Tiglyol DH 1.39 Arnifolins Arnifolin 1.54 11 α,13-dihydroarnifolins 11 α,13-dihydroarnifolin 2.31 Chamissonolides Chamissonolide 0.48 Sum 7.61
Douglas et al. (2004) investigated the levels of sesquiterpene lactones in different stages of flower development of A. montana . They found levels to increase from bud (5.12 mg/ml) to withered (9.43 mg/ml) although the pharmacopoeias recommend harvesting at the partially open stage. This study follows the pharmacopeia recommendations in terms of harvesting but also investigates the levels of compounds in different floral stages of the Orkney grown material both for
A. montana and A. chamissonis .
It is important for any analytical procedure that the limit of detection (LOD), the point at which analysis is just feasible and the limit of quantification (LOQ) the concentration at which quantitative results can be reported with confidence, are determined. LOD can either be determined statistically by measuring repeated blank samples, or by measuring progressively more dilute concentrations of analyte. The latter method has been found to produce more realistic values
(Armbruster et al.1994) and so will be adopted in this study.
82 2.2 Methodology
2.2.1 Flower material
Material used for method validation was harvested from the yield trials (AM2 and
AC2, Section 3.2.1) in the summer of 2005. For the identification and quantification of compounds in the solvent extract, commercial dried A. montana flower material purchased from the Organic Herb Trading Company (2007), was extracted (n=3) alongside the Orcadian A. montana flowers for comparison. The commercial material had France marked as source of origin, although it was stressed by the supplier that this could include Italian sourced flowers depending upon season.
Unfortunately no further information was available from the suppliers, though a voucher sample was taken and stored at ca -20°C. Th e flowers had been dried with warm air and were described as having a shelf life of 2.5 years but were extracted and analysed one month after arrival.
For identification and quantification of compounds in A. chamissonis material harvested in the summer of 2006 from the AC2 trial was used (see section 3.1.2).
To determine any variation in sesquiterpene lactone content of A. montana and A. chamissonis, flowers at different stages of floral development (n=2 for each stage), were harvested. The three stages were: buds (not opened), ‘rays to ½ disc’ (where up to half of the disc flowers were open) and fully open (over) after Douglas et al
(2004). A. montana flowers were taken from AM4 (a density trial that was not part of this study) and the A. chamissonis flowers were taken from AC2, both in the summer of 2006.
83 2.2.2 Extraction methods
2.2.2.1 Hydrodistillation
Essential oil was extracted from the 2004 field season material. A. chamissonis flowers were harvested from AC2 (see section 3.1.2) while the A. montana flowers, roots and stems were from a trial area that was no longer going to be used but that otherwise had been treated as per AM2. The low level of oil produced limited the number of possible repeat analyses, but where possible samples were both extracted and run twice and the average of the four sets of results taken.
All flower material harvested in Orkney was dried at 40°C for four days, the first day in a fan assisted Venticell drying oven (model 111, Einrichtungen GmbH
Germany) which removed the majority of the moisture. In the following days the material was dried in a non fan assisted Electrolux drying oven (Model 234,
Sweden) till constant weight. The dried material was then ground in a Bosch KM13 grinder (Krups, Germany) and frozen in individual containers for later analysis.
Approximately 20 g of material was weighed accurately into a 1000 ml round bottomed flask (Quickfit). For some plots less than 20 g of material was available and in such cases the total available was accurately weighed and used instead.
500 ml of distilled water was added to the dried material and the flask attached to a Clevenger (aka Cocking and Middleton) hydro distillation apparatus (Fig 2.7) over a heating mantle (Electrothermal MV 2403). Care was taken to ensure that enough water was added to soak all of the flower material. The amount of oil
84 extracted after 3 hours (after Roki et al. 2001) was noted, then removed and stored at c.a -18°C prior to analysis.
Attempts were made to distill fresh flower material but this proved not to be possible as an emulsion was formed and no clear oil layer was obtained.
Considerably more oil was obtainable from the roots of A. montana as found by
Bomme & Daniel (1994). However, as the roots are normally used for homeopathy (Hamilton, 1852) and are a less sustainable source of extract compared to the flowers, (which can be harvested from the same plant in successive years), it was felt further investigation was beyond the realm of this study and so only one profile was determined.
Figure 2.7 - Clevenger hydro-distillation apparatus
85 2.2.2.2 Solvent extraction
Ground flower material was combined with 30 ml of MeOH/H 2O (1:1 v/v) in centrifuge tubes, shaken at 100rpm on a shaking platform for 30 minutes, centrifuged at 1700g in a IEC Centra 3E centrifuge (IEC, Dunstable, Bedfordshire,
UK) and filtered through Whatman number 1 filter paper. Extraction was repeated a further 2 times and the samples were then concentrated after ‘spiking’ with 1 ml of 1.00 mg/ml santonin (Fig. 2.8) standard solution (in 50:50 MeOH/H 2O). This was added during the solvent extraction process so it could be used both as a standard and internal standard. As a known amount was added, if less than the amount was determined to be present in the analysis, then it would be known that some of the extract had been lost and that the material should be extracted again.
Santonin was used as an internal standard as it is chemically very similar to the sesquiterpene lactones in Arnica, is readily available and elutes before all the
Arnica sesquiterpene lactones. Hence the internal standard was used both as a reference compound and to determine any sample loss (Wagner & Merfort, 2007).
H
O
O
O
Figure 2.8 - Santonin
The concentrated organic compounds in the Arnica extract were then extracted twice from the 50:50 MeOH/H 2O solution, with 50 ml DCM in separating funnels.
86 The combined extracts were then concentrated to dryness, resuspended in 2 ml
EtOAc and filtered with membrane filters (Sartorius). Samples were diluted to ensure that they were within the range of the linear portion of the standard curve.
This was normally a 1:4 ratio. All tinctures prepared were stored at ca -20°C.
For peak identification and quantification of the sesquiterpene lactones in A. montana flowers, solvent extraction was carried out as per the above on material from the Orkney sites and also from the commercial source for comparison (n=3).
Flower material was also extracted from the ACM and AC2 plot at different stages of floral development during the 2007 field season (n=2).
2.2.3 GC-MS methodology
The material was analysed both quantitatively and qualitatively by GC MS (Hewlett
Packard 5972 Mass Spectrometer with HP 5890 Series lIe Gas Chromatograph).
Extracts were run on a 30m x 0.25 mm BP5 0.25 µm film column (Agilent), mass spectra were obtained in EI (Electron Impact ionisation) mode (70eV) and scanning was from 35 – 500 amu after Schmidt et al. (1998).
In the case of the essential oil, samples were diluted 10 µl in 990 µl after Roki
(2001) but diluted in ethyl acetate. For analysis the method of Adams (2001) was adopted and his spectra and retention times were used as references. The injector and transfer line were kept at 220°C and 2 40°C respectively. The oven temperature was programmed from 60°C to 246°C at 3° C per minute then held for
10 minutes. Helium carrier gas was used at 1.02 ml/min, injection volume was 1.0
µl.
87
Analysis of the solvent extract was adapted from Leven & Willuhn (1987) and
Wagner & Merfort, (2007). The injector and transfer line were kept at 290°C, respectively, and the oven temperature programmed to increase from 120°C to
270°C at 10°C per minute, then held for 20 minutes. Helium was used as the carrier gas and the flow was 1.07 ml/min. The injection volume was 1.0 µl and was not split. All samples were run twice and the average of the two values taken.
2.2.4 Compound Identification
The components of the essential oil of A. montana and A. chamissonis extract were identified using reference spectra and retention times established by Adams
(2001), by reference to pure standards of key compounds (e.g pinene and limonene for the essential oil extract and santonin for the solvent extract) and by comparisons to the work of Kating et al. (1970), Roki et al. (2001) & Willuhn,
(1972a,b &c). The response of the standard was found to be linear between 0.08 and 0.8 mg/ml and so care was taken that the extracts were within this range of concentrations.
Identification of components was determined via GC-MS analysis, first by the peak position relative to santonin (standard and internal standard) following Willuhn &
Leven (1991), confirmed by the mass spectra, then the ratio of the ions, parents, fragments or both. A sample of pure helenalin was also purchased from
Chromadex (CDX08078-0.5) and used to confirm peak identification.
88 Ions 262 and 246 were present in the case of dihydrohelenalin derivatives and ions 262 and 244 were used to identify helenalin derivatives. The ester group was always immediately apparent as an abundant mass fragment and was easily identifiable for most compounds. Where the ester groups were similar (e.g. for 2- methylbutyryl helenalin and isovaleroyl helenalin) the order of elution was used, again after Willuhn & Leven (1991).
2.2.5 Quantification of sesquiterpene lactones
The peak areas in the GC-MS were determined by multiplying the peak width at half height by the total peak height. The GC-MS calibration factors f(sl/st), were determined by the following equation.
f(sl/st) = Z(st) M(sl) / Z(sl) M(st) = f(sl) / f(st)
Where the standard substance f(st) is equal to 1 and where Z = the number of hydrogen bearing carbon atoms in the molecule (santonin = 10) and M = the molecular weight (santonin = 246) and
f(sl) = 0.04065 M(sl) / Z(sl)
The content of each sesquiterpene lactone c(sl), in the sample (s) was calculated by the equation:
c(sl) = A(sl) c(st) f(sl) / A(st)
89 where c(st) is the amount of santonin (mg) added to the crude extract of the sample and A is the abundance of the sesquiterpene lactone. The calibration factors for the compounds of Arnica montana extract were used as per Willuhn &
Leven (1991), and as per table 2.5.
Table 2.5 – The correction factors used for the calculation of individual sesquiterpene lactone content in A. montana flowers
After Willuhn & Leven, 1991
Molecular No of H Name of SL Correction weight bearing C's Santonin 246 10 N/A Dihydrohelenalin 264 12 0.89 Helenalin 262 11 0.97 Acetyl helenalin 304 12 1.03 Acetyl dihydrohelenalin 306 13 0.96 Isobutyryl helenalin 332 14 0.96 Isobutyryl dihydrohelenalin 334 15 0.91 Methacryl helenalin 330 13 1.03 Methacryl dihydrohelenalin 332 14 0.96 Isovaleroyl helenalin 346 15 0.94 2-methylbutyryl helenalin 346 14 1.00 Isovaleroyl dihydrohelenalin 348 16 0.88 2-methylbutyryl dihydrohelenalin 348 15 0.94 Tigloyl helenalin 344 14 1.00 Tigloyl dihydrohelenalin 346 15 0.94
The calibration factors for the compounds of Arnica chamissonis extract were used as per Leven & Willuhn, 1987 (Table 2.6)
90 Table 2.6 – The correction factors used for the calculation of individual sesquiterpene lactone content in A. chamissonis flowers
After Leven & Willuhn, 1987
No of H Molecular Name of SL bearing Correction weight C's Santonin 246 10 N/A Helenalin 262 11 0.968 Dihydrohelenalin 264 12 0.894 4-O-Acetyl-6- 346 15 0.938 desoxychamissonolide Chamissonolide 324 13 1.013 Arnifolin 362 14 1.051
LOD was determined by analysing a series of dilute santonin standards until the peak area was determined to be three times the background noise. Below this value, analysis was deemed not to be possible. The LOQ is defined as the concentration at which all acceptance criteria are met (Armbruster et al. 1994) and in this study was defined as the concentration in which the signal was ten times the background noise.
A t-test (Genstat) was then used to compare the amounts of each sesquiterpene lactone in commercially available dried A. montana flower material with the Orkney grown dried flowers, while an analysis of variance (Genstat) was used to compare the content of sesquiterpene lactones in different floral stages. Data was plotted to check for normality of distribution and where data was required to be transformed it is described in the results.
91
2.3 Results
2.3.1 Calibration curve
Santonin was dissolved in 50:50 MeOH/H 2O and by serial dilution, a set of standards ranging from 0.001 mg/ml to 1.0 mg/ml were created and analysed as per section 2.2.3 and 2.2.4 (Fig 2.9). This curve demonstrated linearity between
0.08 and 0.8 mg/ml (Fig 2.10) and so care was taken that all samples were diluted so as to be within this range.
8000000 7000000 6000000
5000000
4000000
3000000 2000000
Response(PeakArea) 1000000
0 0.000 0.200 0.400 0.600 0.800 1.000 -1000000 Concentration (mg/ml)
Figure 2.9 - Santonin calibration curve
The lowest standard was run six times to determine the level of variation (Table
2.7) which was deemed to be minimal at 0.002 or 0.02 %.
92
6000000
5000000
4000000
3000000
2000000
ResponseArea) (Peak 1000000
0 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 Concentration (mg/ml)
Figure 2.10 - Santonin standard line
Table 2.7 – Level of variation in repeated analysis of 0.08 mg/ml santonin standard
Calculated Theoretical Sample Deviation Response concentration concentration Number (%) mg/ml mg/ml 1 457136.00 0.078 0.080 -2.36 2 477920.00 0.082 0.080 1.90 3 474330.00 0.081 0.080 1.16 4 475861.00 0.081 0.080 1.48 5 498068.00 0.085 0.080 6.02 6 495750.00 0.084 0.080 5.55 Average 479844.17 0.082 0.080 2.29 Standard 15170.62 0.002 0.000 3.106 deviation
LOD was ascertained by analysing the same series of dilute santonin standards that were used to determine the calibration curve. In the 0.02 mg/ml standard the peak area was determined to be three times the background noise and so this was
93 set as LOD. LOQ was found to be 0.05 mg/ml which was just below the minimum limit of linearity (0.08 mg/ml). As a precaution, the minimum sesquiterpene lactone concentration upon which statistical analysis was conducted was 0.08 mg/ml which was slightly higher than the minimum recommended by Willuhn & Leven
(1995).
2.3.2 Hydrodistillation extraction results
The hydrodistillation technique yielded 0.05 - 0.1 ml oil from 40.0g shredded A. montana flowers, 0.05 -0.1 ml oil from dried and shredded stems and 0.2 ml from
25.5g dried rhizome material. From dried and shredded A. chamissonis flowers
0.03 - 0.05 ml oil was obtained. The rhizome oil had a very pungent smell and was prone to emulsify if left distilling for more than 1 hour. The A. chamissonis oil was more viscous than the A. montana oil and as a result was more difficult to isolate as it tended to stick to the glass. The stem oil was very clear and quite runny.
2.3.3 Results of essential oil analysis
The extracted oil (section 2.2.2.1), was analysed by GC-MS using the methodology described in section 2.2.3 and typical chromatographs are illustrated in Figures 2.11, 2.12 & 2.13, 2.14 and 2.15 with detailed results in table 2.8. The latter shows that no sesquiterpene lactones were found in the essential oil extracts. Although the A. montana flower and stem oil was quite similar in terms of content and amount produced, there were some key differences. The stem oil (Fig
2.12), contained more thymol based compounds, caryophyllene and caryophyllene oxide than the flower oil (Fig 2.11), while the flower oil contained more camphene,
94 phellandrene, limonene, linalool, decanal, bermagotene, pentacosane, and unknown compounds.
The root oil of A. montana (Fig 2.13) also contained phellandrene and limonene, as well as even more of the thymol compounds than the stem oil, however, it contained no caryophyllene or caryophyllene oxide. The A. chamissonis flower oil
(Fig 2.14) on the other hand had considerably more pinene than the other samples
(both alpha and beta), cymene, limonene, decanal, spathulenol, tricosane, tetracosane, pentacosane and heptacosane. It contained similar levels of muurolene to the A. montana flowers but hardly any of the thymol compounds were present. Levels of individual compounds for each oil are displayed in figure 2.15 for easier comparison.
95
Figure 2.11 Chromatograph of the essential oil ext oil ofthe essential 2.11 Figure Chromatograph
ract of
A. montana A.
flowers (see table 2.8) table flowers (see
96
ext oil ofthe essential 2.12 Figure Chromatograph
ract of
montana A.
stems (see table 2.8)stems(see table
97
ext oil essential of the 2.13Figure Chromatograph
ract of
montana A.
roots (see table 2.8)roots (see table
98
Figure 2.14 Chromatograph of the essential oil ext oil of the essential 2.14Figure Chromatograph
ract of
A. chamissonis A.
flowers (see table 2.8)flowers (see table
99 Table 2.8 Compounds identified in the essential oil analysis (AMF – A. montana flowers, AMS – A. montana stems, AMR – A. montana roots and ACF – A. chamissonis flowers.
Percentage Content AMF AMS AMR ACF No. t Compound Formula CAS Mw R oil oil oil oil
1 5.008 Pinene
2 5.342 Camphene C10 H16 79-92-5 136 0.68 0.20 0.39 0.00
3 6.093 Pinene
4 6.857 Phellandrene
5 7.471 Cymene
6 7.611 Limonene C10 H16 138-86-3 136 0.64 0.18 0.33 1.41
7 10.092 Linalool C10 H18 O 78-70-6 154 1.15 0.27 0.00 1.29
8 10.240 Nonanal C9H18 O 124-19-6 142 0.21 0.20 0.00 1.03
9 14.279 Decanal
10 15.498 Thymol, methyl ether C11 H16 O 1076-56-8 164 0.40 2.57 31.31 0.00 8,9-dehydro-thymol-methyl 11 16.088 C H O 162 0.42 4.12 8.63 0.44 ether 11 14
12 17.985 Thymol C10 H14 O 89-83-8 150 0.20 6.69 1.77 0.27
13 22.730 Dodecanal C12 H24 O 112-54.9 184 0.00 0.00 0.00 0.83
14 23.063 Caryophyllene <(E)> C15 H24 87-44-5 204 24.37 29.09 0.00 0.99 8,9-dehydro-4-hydroxy- 15 23.319 C H O 194 0.24 1.58 23.85 0.00 thymol-dimethyl ether 12 18 2 Bergamotene
18 24.425 Humulene
19 24.712 Farnesene <(E)-beta-> C15 H24 18794-84-8 204 1.50 0.70 0.14 0.30
20 25.531 Muurolene
21 25.705 Isobutyl-thymyl ester C14 H20 O2 220 4.52 1.75 3.81 0.00 Isobornyl 2- 22 26.544 C H O 94200-10-9 238 0.75 1.87 0.34 0.00 methylbutanoate 15 26 2
23 29.200 Spathulenol C15 H24 O 6750-60-3 220 0.00 0.00 0.00 4.58
24 29.359 Caryophyllene oxide C15 H24 O 1139-30-6 220 3.52 18.76 0.00 2.91
25 31.832 Desmethoxy encecalin C13 H14 O2 19013-07-1 202 0.00 2.75 0.00 1.12 Eudesma-4(15),7-dien-1- 26 33.194 C H O 000-00-0 220 0.74 3.81 0.00 4.77 beta-ol 15 24
27 38.860 Perhydrofarnesylacetone C18 H36 O 502-69-3 268 2.64 0.26 0.00 1.63 28 44.665 Peak A 2.36 9.12 9.17 0.38 29 47.286 Peak B 1.33 4.13 3.43 0.78 30 48.842 Peak C 2.69 0.00 0.00 0.00 31 50.388 Peak D 2.22 0.00 0.00 0.00 32 51.842 Peak E 4.48 0.00 0.00 0.00
33 52.712 Tricosane
34 55.394 Tetracosane
35 57.974 Pentacosane
36 62.898 Heptacosane
(It was not possible to identify peaks A-E)
100
– AMF
montanaA.
AMS – flowers,
A. montana montana A.
analysis oil FigureEssential 2.15 stems, AMR – AMR stems,
montana A.
- ACF and roots
A. chamissonis chamissonis A.
Table 2.8) (from flowers
101 2.3.4 Identification of compounds in the solvent extract of A. montana
A typical chromatograph for the extract from dried A. montana flowers from Orkney is shown in figure 2.16, with one for extract from commercially available dried flower material is shown in figure 2.17. The solvent extract of the A. montana flowers (section 2.2.2.2), was analysed by GC-MS using the methodology described in section 2.2.3 and as described in section 2.2.4, the peaks were identified by comparison with published analysis of similar extracts (Willuhn &
Leven, 1991: Malarz et al. 1993) and this process is described in sections 2.3.4.1-
2.3.4.14.
Santonin 1 2 3 4 5 6 7 8,9,10,11,12 13 14
Figure 2.16 – Sample chromatograph of the Orcadian A. montana solvent extract
Santonin 1 2 3 4 5 6 7 8,9,10,11,12 13 14
Figure 2.17 – Sample chromatograph of the commercially available A. montana solvent extract
102 2.3.4.1 Peak 1
Figure 2.18 – Retention time of Peak 1
H
O
O O HO
Figure 2.19 - 11 α13-dihydrohelenalin
Figure 2.20 - The mass spectra of peak 1
103 Peak 1 (Fig 2.18) was identified as tenulin, or 11 α13-dihydrohelenalin (C 15 H20 O4)
M+=264 (Fig. 2.19), corresponding to the work of Willuhn and Leven (1991).
Although the parent ion is absent, the mass spectral data contains the ion fragment 44 (CO 2), and both 55 and 124 (Fig 2.20) which appear in all helenalin and dihydrohelenalin spectra. The 231 fragment is characteristic of all the dihydrohelenalin spectra and the 191 fragment hypothesised to result from the cleavage of the lactone –hemiketal system to form the ion in Figure 2.21.
H H
O
O O 191 O + O O
H
Figure 2.21 - Formation of the 191 fragment
As proposed by Tsai et al. 1969
2.3.4.2 Peak 2
Figure 2.22 - Retention time of Peak 2
104 H
O
O O HO
Figure 2.23 – Helenalin
Figure 2.24 - The mass spectra of peak 2
Peak 2 (Fig 2.22) was identified as helenalin (M +=262, although the parent ion was not present) (Fig 2.23), both from running pure helenalin standard parallel to this analysis and by comparing retention time data from Willuhn and Leven (1991). It has a similar profile to that of peak 1 in that it contains the 44, 55 and 124 fragments (Fig 2.24), but it also has the 229 fragment characteristic of all helenalin spectra. The 96 fragment was believed to result from the fragmentation route that produced m in Figure 2.3 (Tsai et al. 1969).
105 2.3.4.3 Peak 3
Figure 2.25 – Retention time of Peak 3
Figure 2.26 - The mass spectra of peak 3
Peak 3 (Fig 2.25) contains a high proportion of the 43 ion (Fig 2.26) which is indicative of an acetyl ester group (Fig 2.27), and would correspond with results of
Malarz et al. (1993) who also found the 262, 244 and 229 ions in their analysis of acetyl helenalin. A prominent peak at m/e 244 was also found for acetyl helenalin by Tsai et al. (1969) which was believed to be due to the removal of acetic acid
(Fig 2.25). Together, this evidence indicates that peak 3 corresponds to acetyl helenalin (C 17 H20 O5).
106
O
CH3
Figure 2.27 - The acetyl group
244 O O
O O HO O O H C CO H2
Figure 2.28 - The removal of acetic acid from acetyl helenalin
2.3.4.4 Peak 4
Figure 2.29 – Retention time of Peak 4
107
Figure 2.30 - The mass spectra of peak 4
Peak 4 (Fig 2.29) also has the 43 ion (Fig 2.30) corresponding to an acetyl group
(Fig 2.27), but in this case has a 246 ion instead of 244. This is believed to result from a similar mechanism to Fig 2.25, but with the two extra protons of dihydrohelenalin (Fig 2.28). Malarz et al. (1993) also found the mass spectra data for this compound to contain 264, 246, 231 and 43. It is hence believed that this peak corresponds to acetyl dihydrohelenalin (C 17 H22 O5).
246 O O
O O HO O O H C CO H2
Figure 2.31 - The removal of acetic acid from acetyl dihydrohelenalin
108 2.3.4.5 Peak 5
Figure 2.32 – Retention time of Peak 5
Figure 2.33 - The mass spectra of peak 5
Peak 5 (Fig 2.32) has a large peak at 43 and also at 71 (Fig 2.33) which would correspond to the isobutyryl group (Fig 2.34). It also has the 244 and 262 ions which correspond to helenalin ester, and a 332 which would suggest the molecular ion of isobutyryl helenalin (C 19 H24 O5). A 332 ion could also correspond to methacyrl dihydrohelenalin and 2 methylbutyryl helenalin, but in this case as there
109 is no evidence of the corresponding 69 or 85 fragments it can be concluded this peak is that of isobutyryl helenalin.
O
CH3
CH3
Figure 2.34 - The isobutyryl group
2.3.4.6 Peak 6
Figure 2.35 – Retention time of Peak 6
Peak 6 (Fig 2.35) contains the 231, 246 and the 264 ions (Fig 2.36) which indicates that it is a dihydrohelenalin ester. It also contains the 71 ion which indicates an isobutyryl ester group (Fig 2.34). Although the parent ion (M + 334) was not present, it can be concluded that this peak corresponds to isobutyryl dihydrohelenalin (C 19 H26 O5).
110
Figure 2.36 - The mass spectra of peak 6
2.3.4.7 Peak 7
Figure 2.37 – Retention time of peak 7
The 229 and 244 ions of peak 7 (Fig 2.37), indicate it is a helenalin ester. The mass spectra data corresponds to the 330, 261, 244, 229 and 69 found by Malarz et al. (1993) in their analysis of methacryl helenalin. The 69 ion (Fig 2.38) would correspond to the methacryl ester (Fig 2.39) and 330 to the parent ion. Together, this evidence confirms that this peak corresponds to methacryl helenalin
(C 19 H22 O5).
111
Figure 2.38 - The mass spectra of peak 7
O
CH2
CH3
Figure 2.39 - The methacyrl group
2.3.4.8 Peak 8
Figure 2.40 – Retention time of peak 8
112
Figure 2.41 - The mass spectra of peak 8
Peak 8 (Fig 2.40), contains the 263, 246, 231 and 69 ions (Fig 2.41) found by
Malarz et al. (1993) for methacryl dihydrohelenalin (C 19 H24 O5). This, combined with its retention time leads to the conclusion that peak 8 is methacryl dihydrohelenalin.
2.3.4.9 Peak 9
Figure 2.42 – Retention time of peak 9
113
Figure 2.43 - The mass spectra of peak 9
Peak 9 (Fig 2.42), contains the 229, 244 and 262 ions (Fig 2.43) which indicate a helenalin ester and has the 85 ion which could correspond to either 2- methylbutyryl (Fig 2.44), or isovaleroyl (Fig 2.45). However, the 57 peak is larger than the 85, which implies a connection in the ester group that is subject to breakage. The 2- methylbutyryl is less likely to fragment than the isovaleroyl particularly along the red marked line to produce the fragment 57, and the green line to produce the fragment 41 and so we can conclude that this peak is isovaleroyl helenalin (C 20 H26 O5).
O
CH3
CH3
Figure 2.44 - The 2-methylbutyryl group
114 O
C
Figure 2.45 - The isovaleroyl group
2.3.4.10 Peak 10
Figure 2.46 – Retention time of peak 10
Figure 2.47 - The mass spectra of peak 10
115 Peak 10 (Fig 2.46), contains the 229, 244 and 262 ions (Fig 2.47) which indicate that it is a helenalin ester. It also has 57 and the 85 ion as well as the likely molecular ion of 346. In this case it has a larger 85 peak than 57, and so it is concluded that this peak corresponds to 2-methylbutyryl helenalin.
2.3.4.11 Peak 11
Figure 2.48 – Retention time of peak 11
Figure 2.49 - The mass spectra of peak 11
116 Peak 11 (Fig 2.48), contains the 231, 246 and 264 ions (Fig 2.49) which indicate that it is a dihydrohelenalin ester. In this case, as the 57 is larger than the 85 ion, it is concluded that this corresponds to an isovaleroyl ester group and isovaleroyl dihydrohelenalin.
2.3.4.12 Peak 12
Figure 2.50 – Retention time of peak 12
Figure 2.51 - The mass spectra of peak 12
117 Peak 12 (Fig 2.50) also contains the 231 and 244 ion (Fig 2.51) which is indicative of a helenalin ester, as are the 57, 69 and 85 ions. However, in this case, as the
85 is smaller than the 57 it is predicted that this corresponds to 2-methylbutryl dihydrohelenalin.
2.3.4.13 Peak 13
Figure 2.52 – Retention time of peak 13
Figure 2.53 - The mass spectra of peak 13
118 Peak 13 (Fig 2.52), contains the 229 and 244 ion (Fig 2.53) which suggests a helenalin ester. It also has the 83 ion which in this case is predicted to correspond to a tigloyl ester (Fig 2.54), and a 344 ion which is likely to correspond to the molecular ion. It was hence concluded that this peak corresponded to tigloyl helenalin (C 20 H24 O5).
O
CH3
CH3
Figure 2.54 - The tigloyl fragment
2.3.4.14 Peak 14
Figure 2.55 – Retention time of peak 14
Peak 14 (Fig 2.55) contains the 246 ion (Fig 2.56) which indicates a dihydrohelenalin ester, as well as a fragment corresponding to m/e 263. The presence of the 83 ion corresponds to a tigloyl ester group (Fig 2.54). It can be concluded that this peak corresponded to tigloyl dihydrohelenalin (C 20 H26 O5).
119
Figure 2.56 - The mass spectra of peak 14
2.3.5 Quantitative analysis of A. montana solvent extract
2.3.5.1 Commercial and Orcadian extract
Quantification of all the above peaks was carried out as described in section 2.2.5 and the individual sesquiterpene lactone content for extract from both the commercial and Orcadian material is presented in Table 2.9 and Figure 2.57.
Whilst there was some variation in the content of the individual sesquiterpene lactones (table 2.9 and Fig. 2.57) this was often in values below the LOQ and there was no significant difference in total content (t(4)=0.8, p=0.49) (Table 2.9 and Fig. 2.58). The main differences between the two sources of flowers seem to be that there are significantly more of the helenalin esters in the Orcadian extract
(t(4)=5.8. p=0.004), and more of the dihydrohelenalin esters (t(2)=25.6, p=0.001), in the commercial flower extract.
120 Table 2.9 - Sesquiterpene lactone content of extract from commercial and Orcadian A. montana flowers
Orkney Commercial Sesquiterpene lactones t= p= (mg/ml) (mg/ml) Dihydrohelenalin 0.03 0.02 4.45 0.011 Helenalin 0.05 0.01 15.47 <0.001 Acetyl helenalin 1.24 0.59 18.85 0.002 Acetyl dihydrohelenalin 0.05 0.20 -81.04 <0.001 Isobutyryl helenalin 1.37 1.53 -2.25 0.088 Isobutyryl dihydrohelenalin 0.03 0.47 -45.07 0.001 Methacryl helenalin 1.45 0.98 8.66 0.001 Methacryl dihydrohelenalin 0.04 0.35 -16.33 0.003 Isovaleroyl helenalin 0.99 0.99 0 1 2-methylbutyryl helenalin 0.41 0.21 6.12 0.004 Isovaleroyl dihydrohelenalin 0.01 0.13 -16.88 0.001 2-methylbutyryl dihydrohelenalin 0.01 0.09 -7.15 0.017 Tigloyl helenalin 0.58 0.38 4.56 0.010 Tigloyl dihydrohelenalin 0.02 0.15 -13.35 0.005 Total helenalin and helenalin 6.10 4.69 5.8 0.004 esters Total dihydrohelenalins and 0.19 1.40 25.6 0.001 dihydrohelenalin esters
Total sesquiterpene lactones 6.29 6.08 0.8 0.49
Yellow corresponds to significant difference and italic values are below the LOQ
121
Figure 2.57 - Graph of the sesquiterpene lactone cothe lactone2.57 of Figure - sesquiterpene Graph Total (mg/ml)
D ih y d 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 r o-hele nal
in *
Hele Acet nal *
in Acety yl hel l dihydro-he enal in * *
Is l obutyryl helenal Is ob in * utyryl dihydro-helena
enal in
M et * indicates significant difference between two valu two between difference significant * indicates M hacry l ethacr in *
l hel yl di enal hydro- in
hel * Isovaleroyl helenalin lactones Sesquiterpene ena 2-methylbutyryl helenalinlin *
ntent of extract from commercial andOrcadianntent of fromextract commercial
Iso val eroyl dihyd 2-m
et hyl buty r ohelenalin *
ryl di
Commercial (mg/ml) average
hy *
dr ohel ena Tot Tigloyl helenalin lin * al T Tigloyl dihydrohelenalin di o hydrohel tal helenalin * es ena li an ns d helenal
and dihyd *
Orkney average in es T rohelenalin ot * ters
al sesqui
ter est
p ers *
ene
lact
A. montanaA. ones
flowers
122
7.00 Commercial average (mg/ml) Orkney average
6.00
5.00
4.00
3.00 Total (mg/ml)
2.00
1.00
0.00 Total helenalin and helenalin esters Total dihydrohelenalins and Total sesquiterpene lactones dihydrohelenalin esters Content
Figure 2.58 - Total sesquiterpene lactone content of commercial and Orcadian A. montana flower extract.
The variation in content of the two extracts could be due to where the material was grown, or it could be the result of an agronomic or post harvesting treatment. Due to the secretive nature of this industry, information on either aspect was unfortunately not available upon request. In addition, the country of origin was altered (from Italy, to France, to both Italy and France) when a certificate of production was requested from the source. It was claimed that country of origin varied due to commercial demand, so it may be that flowers of the Spanish chemotype have been included in this batch. As a result, the only conclusion that can safely be drawn, is that the Orcadian extract contains a higher proportion of helenalin esters (which are believed to be the more active constituents) than this
123 commercially available source, but has similar levels of sesquiterpene lactones overall. More analyses on a number of different sources are required before definitive statements on the quality of this material compared to that available on the commercial market can be made.
While the sesquiterpene lactones in both the commercial and the Orcadian extract were identifiable, many were not accurately quantifiable due to a high signal/noise ratio and because they were below the linear part of the santonin calibration line.
The limit of detection was set at 0.02 mg/ml as this was approximately the background noise level whilst the limit of quantification was 0.08 mg/ml.
Sesquiterpene lactones below this were described as being below the limit of quantification (LOQ) and are identified as such in italics in table 2.10. It is apparent that all of the compounds present in quantities above the LOQ are helenalin esters, which is in line with the findings of Willuhn & Leven (1991) and Douglas et al. (2004).
2.3.5.2 Extract from different floral stages
In order to determine whether the types and amounts of sesquiterpene lactones produced by the flowers changed during development, A. montana flowers at different stages (section 2.2.1), were extracted as per section 2.2.2.2 and the results presented in table 2.10 and figures 2.59 & 2.60.
124 Table 2.10 - Sesquiterpene lactone content in the different stages of floral development in A. montana
Partially Sesquiterpene lactones Over open Bud F= p= (mg/ml) (mg/ml) (mg/ml)
Dihydrohelenalin 0.00 0.00 0.01 10.06 0.012 Helenalin 0.01 0.02 0.01 0.93 0.445 Acetyl helenalin 1.04 0.84 0.49 5.01 0.053 Acetyl dihydrohelenalin 0.08 0.06 0.13 18.34 0.003 Isobutyryl helenalin 1.66 1.47 0.69 12.18 0.008 Isobutyryl dihydrohelenalin 0.06 0.05 0.07 0.47 0.647 Methacryl helenalin 1.61 0.99 0.69 7.18 0.026 Methacryl dihydrohelenalin 0.08 0.06 0.17 10.80 0.010 Isovaleroyl helenalin 1.13 0.95 0.48 5.94 0.038 2-methylbutyryl helenalin 0.46 0.32 0.21 2.16 0.197 Isovaleroyl dihydrohelenalin 0.01 0.01 0.02 1.01 0.420 2-methylbutyryl dihydrohelenalin 0.02 0.03 0.03 0.92 0.448 Tigloyl helenalin 0.53 0.38 0.36 0.82 0.484 Tigloyl dihydrohelenalin 0.02 0.03 0.06 15.92 0.004 Total helenalin and helenalin 6.46 4.97 2.93 5.88 0.039 esters Total dihydrohelenalins and 0.26 0.25 0.49 10.93 0.01 dihydrohelenalin esters 6.72 5.22 3.42 -4.64 0.061 Total sesquiterpene lactones
Over = fully open over mature flowers, partial = rays open to half disc, bud = bud stage flowers.Yellow indicates a significant difference between the values in each stage
Although the trend was for the individual sesquiterpene lactones to be of a higher concentration in the over flowers than the partially open flowers, the difference was not significant overall (F=-4.64, p=0.061). However, in the buds there were significantly lower levels of helenalin esters (F=5.88, p=0.039) and higher amounts of dihydrohelenalin esters (F=10.93, p=0.01) (Table 2.10 and Fig. 2.60).
125
Amount (mg/ml) Over = fully open over mature flowers, partial = ra partial flowers, mature over open fully = Over D 0.00 0.50 1.00 1.50 2.00 2.50 ihyd
ro hel Figure 2.59 - Sesquiterpene lactones content of 2.59Figure - Sesquiterpene lactones enal
in *
H ele na lin
Ac e tyl h e l enal in Acety
l dihydrohel * Over average (mg/ml) average Over en alin Isob ut * yryl h elen Is o butyryl dih al in * ys open to half disc, bud = bud stage flowers * ind flowers stage = bud bud disc, half open to ys ydrohel
e nali n M ethac * lactones Sesquiterpene ry l he A. montana A.
three valuesthree Me le thac nal in Partially open (mg/ml) average ry l di hy *
drohe len al in Isov flowers in different stages of floral development ofdevelopment floral stages different flowersin al * eroy l h el e 2-methylbutyryl helenalinn al in *
Is ov ale ro yl di h ydrohe icates significant difference between the the between difference significant icates 2 -m eth yl len buty al in Bud (mg/ml) average ry l di hy drohe
len al in Tigl oy l h ele nal Ti g in loy l dihy d ro helena
lin
*
126 Schmidt et al. (1998) suggest the presence of a hydrogenase system that converts sesquiterpene lactones from the hydro to the dihydro type and suggest that further investigation into this pathway would be of interest. This study found levels of dihydrohelenalin esters to be higher in A. montana buds than in the open flowers and this could be an indicator of such a system.
It should be noted that the average yield for the partially open flowers was lower than was found in section 2.3.5.1. The material for the comparison of commercially available and Orkney grown material was grown in 2007, whilst material for the comparison of buds, over and partially open material was grown in 2006.
10.00 Over average (mg/ml) Partially open average (mg/ml) Bud average (mg/ml) 9.00
8.00
7.00
6.00
5.00
* Amount (mg/ml) 4.00 *
3.00
2.00 * 1.00
0.00 Total helenalin and helenalin esters Total dihydrohelenalins and Total sesquiterpene lactones dihydrohelenalin esters Sesquiterpene lactone type Figure 2.60 - Graph of the total sesquiterpene lactone content of A. montana flowers at different stages of development
* indicates significant difference between the three values
127
2.3.6 Solvent extract compound identification for A. chamissonis
The solvent extract of the A. chamissonis flowers (see section 2.2.2.2), was analysed by GC-MS using the methodology described in section 2.2.3 and the peaks were identified as per section 2.2.4. A typical chromatograph for the extracts is shown in figure 2.61. As well as peaks for dihydrohelenalin (peak 1) and helenalin (peak 2) (see 2.3.4.1 and 2.3.4.2) additional peaks were identified and these are described in sections 2.3.6.1-2.3.6.3.
15 16 17
Figure 2.61 - Chromatograph of the A. chamissonis solvent extract
2.3.6.1 Peak 15
Figure 2.62 – Retention time of peak 15
128
Figure 2.63 - The mass spectra of peak 15
Peak 15 (Fig 2.62) of the A. chamissonis extract contains the 55 and 124 ions (Fig
2.63) characteristic of all the sesquiterpene lactones (Tsai et al. 1969). Although the 308 parent ion is missing in this spectra, the following MS m/z: 248, 230, 215,
159, 119, 108, 107 and 43 were all present as they were for Willuhn et al. (1983) along with a 230 and 248 ion believed to result from the fragmentation outlined in
Figure 2.64. Together, this indicates 4-O-Acetyl-6-desoxychamissonolide (Fig
2.65).
129 O
H H3C C O
O O O O H C C O 3 O H3C C O O H H 291
O O
HO O H O H 248 230
Figure 2.64 – The fragmentation of 4-O-Acetyl-6-desoxychamissonolide
O
H H3C C O
O O
H3C C O O H
Figure 2.65 – 4-O-Acetyl-6-desoxychamissonolide
130 2.3.6.2 Peak 16
Figure 2.66 – Retention time of peak 16
Figure 2.67 – The mass spectra of peak 16
Peak 16 (Fig 2.66) of the A. chamissonis extract contains the 264, 246 and 228 ions (Fig 2.67) which indicates the presence of chamissonolide (Fig 2.68). The 264 corresponds to the loss of an acetate group and hydrogen (Fig 2.69), the 246 to the subsequent loss of water, whilst the 228 in turn corresponds to the loss of water from the 246 (Fig 2.69).
131 O
H C C O 3 H
O
HO O HO
Figure 2.68 – Chamissonolide
O
H H3C C O
O O
O O HO O HO HO 264
O
O O
HO O 228 246 Figure 2.69 – Fragmentations of chamissonolide
132 2.3.6.3 Peak 17
Figure 2.70 – Retention time of Peak 17
Figure 2.71 – The mass spectra of peak 17
Peak 17 (Fig 2.70) of the A. chamissonis extract contains a 262 and 244 ion (Fig
2.71) which indicates the presence of arnifolin (Fig. 2.72). The 262 corresponds to the loss of the tiglate group (the 83 ion) and one OH group (362-83-17 = 262) the
244 corresponds to the loss of water from the 246 (Fig 2.73). The mass spectra is also very similar to that determined by Zakharov et al. (1971).
133
HO H
O
O O O
C O
C CH3
HC CH3
Figure 2.72 – Arnifolin
HO H HO H
O O
O O O O O
C O 262
C CH3
HC CH3
H
O
O O
244
Figure 2.73 – Fragmentation of Arnifolin
134 2.3.7 Quantitative analysis of A. chamissonis solvent extract
2.3.7.1 A. chamissonis extract
The relative amount of each peak was calculated as described in section 2.2.5.
The results for an analysis of A. chamissonis flowers (n=6) are shown in Table
2.11 and Figure 2.74. It was decided, as for A. montana, that sesquiterpene
lactones too close to the background noise level and below the linear part of the
santonin calibration line should be classified as below the LOQ. The limit was
judged to be at the 0.08 mg/ml level. As was the case for the A. montana extract,
the compounds present in quantities above the LOQ are helenalin type esters,
which is consistent with the findings of Leven & Willuhn (1987) and Willuhn &
Leven (1995). By far the most significant sesquiterpene lactone in the extract was
chamissonolide with an average content of over 1 mg/ml. This was not consistent
with the findings of the aforementioned authors, which could be due to agronomic
conditions, or the source of the original seed. As a result, both factors were
investigated.
Table 2.11 – A. chamissonis extract
No of H More Peak Molecular Correction Amount Compound bearing Response than No. weight factor (mg/ml) C's LOQ? Santonin 246 10 1156450 1.00 N/A
1 Dihydrohelenalin 264 12 0 0.89 0.00 X
2 Helenalin 262 11 2684 0.97 0.00 X 4-O-Acetyl-6- 3 desoxy- 350 14 89936 1.02 0.08 chamissonolide 4 Chamissonolide 324 13 1558075 1.01 1.36
5 Arnifolin 362 13 250221 1.13 0.24
135 1.6
1.4
1.2
1
0.8 Amount (mg/ml) Amount 0.6
0.4
0.2
0 Dihydrohelenalin Helenalin 4-O-Acetyl-6- Chamissonolide Arnifolin desoxychamissonolide Sesquiterpene lactone
Figure 2.74 - Sesquiterpene lactone content of A. chamissonis flowers (n=6)
Bars represent standard deviation
2.3.7.2 Extract from different floral stages
In order to determine whether the types and amounts of sesquiterpene lactones produced by the flowers changed during development, analysis was also run on different stages of A. chamissonis flowers.
Because the A. chamissonis flowers are smaller than A. montana, it was thought likely that the harvesting method most suited to this size of flowers would also result in more stem and leaf material being present in the final extract. Hence an
136 analysis of extracted leaves and stems (n=3) was also conducted. The results are presented in table 2.12 and figure 2.75.
Although helenalin and dihydrohelenalin had been found previously in A. chamissonis extract, neither sesquiterpene lactone was found in this analysis.
Sesquiterpene lactone content of the leaves and stem were very obviously different from the floral parts and so were excluded from statistical analysis.
However, even with this data removed, the content of each sesquiterpene lactone varied significantly with stage of flowering (Table 2.12 and Fig 2.75), with the highest proportion being found in partially open flowers. This contrasted with the results for A. montana (section 2.3.5.2) which found the highest level in flowers that had gone beyond the recommended harvesting stage.
Table 2.12 - Results for A. chamissonis floral stages
Partially Over Bud Stem Leaves open F= p= average average average* average* average 4-O-Acetyl-6-desoxy- 0.06 0.12 0.02 0.01 0.00 39.82 <.001 chamissonolide
1.10 1.47 0.37 0.00 0.07 221.36 <.001 Chamissonolide
0.18 0.26 0.07 0.00 0.12 39.87 <.001 Arnifolin
Total sesquiterpene 1.35 1.85 0.46 0.01 0.20 172.86 <.001 lactones
*excluded from ANOVA, italics indicate values below LOQ
137
2.50
Over average Partially open average Bud average Stem average Leaves average
2.00
1.50
Amount (mg/ml) 1.00
0.50
0.00 4-O-Acetyl-6- Chamissonolide Arnifolin Total sesquiterpene lactones desoxychamissonolide Sesquiterpene lactones
Figure 2.75 – Levels of sesquiterpene lactones in different stages of floral development and of the leaves and stems in A. chamissonis
Bars represent standard deviation
2.4 Discussion of the qualitative and quantitative extract results
The aim of this chapter was to determine the quality and quantity of the extracts of
Arnica grown in Orkney and to compare it to commercially available sources and
138 those described in the literature. The analytical methodologies developed for both essential oil and solvent extract proved to be effective for both A. montana and A. chamissonis material grown in the Orkney field sites, and for commercial material.
Overall, the content of the extracts were similar to that found in the literature, although some qualitative and quantitative differences were found within the extracts.
The amount of essential oil produced was low, ranging from 0.03 – 0.05 ml for A . chamissonis flowers, 0.05-0.1 ml for the A. montana flowers and stems, 0.2 ml for the roots per 40 g dried material. The small amount of oil available for each type of material led to some sample loss and difficulties with repetitions of analysis.
However, clear variations in the content of the oils were apparent.
Roki et al. (2001) who introduced A. chamissonis to the Sera mountain area of
Serbia, found that the essential oil of A. chamissonis consisted mostly of n- alkanes, sesquiterpenoids and monoterpenoids whilst that of A. montana contained mostly fatty acids, the levels of which were much lower in A. chamissonis. This was also found in this analysis (Table 2.8), where the total n- alkane content of A. montana floral essential oil was found to be 24 %, compared to the 45 % content of A. chamissonis. Roki et al. (2001) also found in their analysis of A. chamissonis essential oil that the most abundant n-alkane was n- pentacosane with a percentage content of 9.5%. In A. chamissonis oil produced from Orkney flowers, the same compound was found at much higher levels
(22.9%) and was by far the dominant constituent of the oil, with n-tricosane
(13.6%) also present at levels higher than the 6.5% found by Roki et al. (2001).
139
Levels of n-heptacosane were also found to be higher in this analysis than in the literature (5.2% vs 3.2%) although the percentage content of n-tetracosane at
1.70% was similar. The Roki et al. (2001) study found higher levels of caryophyllene oxide in A. chamissonis oil (5.5% vs 2.8%) but lower levels of spathulenol (3.6% vs 4.33%). Of the other compounds found to be present, nonanal was high at 9.4% as was gamma-muurolene at 7.3% and methyl gamma
Ionone at 7.49%. Of the monoterpenes, alpha-pinene was a fairly strong constituent at 8.25%, while p-cymene was found at levels of 2.48%. The high percentage of paraffins explains the viscosity of the extracted A. chamissonis oil and its tendency to stick to the glass as the oil cooled.
In comparison to the A. chamissonis essential oil, the oil produced from the flowers of A. montana contained fewer monoterpenes and lower amounts of each
(Table 2.8). No alpha-pinene was detected, compared to the 8.87% found in A. chamissonis and although there was more phellandrene detected in A. montana
(0.9%) than in A. chamissonis (0.3%), there was more limonene and para-cymene in the A. chamissonis oil. However, there was considerably more caryophyllene
(24.4%) in the A. montana oil compared to just 1% in the A. chamissonis oil, as well as more muurolene at 10.0% compared to 7.9% in A. chamissonis.
The oil produced from A. montana stems was also high in caryophyllene at 29.1%, caryophyllene oxide (18.8%) and relatively high in thymol (6.7%) and thymol methyl ethers. However the stems were very low in n-alkanes compared to the
140 floral oil which explains the lower viscosity of this oil compared to that extracted from the flowers.
Caryophyllene has been shown to have anti-inflammatory properties, as does caryophyllene oxide (Chaven et al. 2010) and humulene (Fernandes et al. 2007) both of which were present in higher amounts in A. montana than in A. chamissonis oil. Both caryophyllene and caryophyllene oxide are also insect repellents (Hubbell et al. 1984) and antifungal agents (Hubbell et al. 1983), with caryophyllene oxide reported as 20 times more effective in the assay than the caryophyllene precursor (Hubbell et al. 1984).
Thymol is antiseptic and antifungal compound (Kordali et al. 2008) which suggests that the stem material could be included in essential oil material along with the flowers without too much of a detrimental effect, depending of course, upon the intended use of the oil.
The greatest amount of thymol however, was to be found in the roots. Willuhn
(1972c) found the main component (46%) of A. montana root oil to be 4-hydroxy- thymol-dimethyl ether and although a high percentage content of this compound was found in this analysis (23.8%), the content of thymol-methyl ether was higher in this study at 31.3% compared to 12% of total content found by Willuhn. In the same study Willuhn also found isobutyl-thymol ester at levels of 8%, whereas in this analysis it was found to be 3.8%.
141 Sesquiterpene lactones were absent from both oils (table 2.8), which could mean that the high temperatures caused a Diels-Alder reaction to produce compounds which were unidentifiable in the resulting mass spectral fragments. They could be the unknown peaks A-E (No’s 28-32 in table 2.8), but as the heating process involved in extracting the oil would destroy the structure of the compounds and hence their utility, this will not be investigated further. Where sesquiterpene lactones are the compounds of interest, solvent extraction is the more common form of extraction (DAB 7, 1964 and DAB 10, 1991).
It has been suggested that increased solar radiation at high altitudes increases the levels of sesquiterpene lactones (Körner, 1999) so hypothetically the longer day length in Orkney could have led to increased levels of sesquiterpene lactones.
However, this study found that the average level of sesquiterpene lactones were not substantially higher than those found in commercial material (table 2.9) or those found by other studies (e.g Willuhn & Leven, 1991), which agrees with the findings of Spitaler et al. (2006), which was that plants exposed to higher solar radiation did not have higher levels of sesquiterpene lactones.
The average sesquiterpene lactone content of 5.2 mg/ml (0.52% w/w) of A. montana flowers grown in 2006 and 6.3 mg/ml (0.63% w/w) found in material grown a year later, are both above the 0.40% w/w recommended for topical Arnica preparations (Willuhn & Leven, 1995). Although the levels may vary seasonally, and are fairly close to the minimum level, the A. montana flowers produced in
Orkney can be claimed to be of high enough a standard for the preparation of herbal tincture. Any agronomic regime that increased this level would, however,
142 greatly increase the confidence in such a statement and this will be investigated in subsequent chapters.
Klaas et al. (2002) demonstrated that the inhibitory effect of helenalin, 11 α-13- dihydrohelenalin and their derivative esters on the inhibition of inflammation was dependent upon the ester groups of the sesquiterpene lactones. Those with unsaturated acyl moieties such as methacrylate and tiglinate were found to display greater activity than the acetate derivatives. Extracts of the Orkney flowers were found to contain significantly higher levels of these compounds than the commercially available source. In the same study, acetyl dihydrohelenalin (which was also found at slightly higher levels in Orcadian flower material) was found to moderately induce apoptosis in Jurkat T-cells.
Although the quality of A. montana material produced in Orkney is therefore of a quality suitable for the commercial market, sesquiterpene lactones in A. montana have been shown to increase with development from buds until the petals withered
(Douglas et al. 2004). This was found to also be the case for plants grown in
Orkney as when the flowers were left till they were fully open, the content of sesquiterpene lactones increased to almost 0.7% w/w. It may hence be the case that if harvesting is left till later than the ‘partially open’ stage that the pharmacopeia recommends (DAB10, 1991) the yield of the active compounds within the extract produced could be improved.
This study also found levels of dihydrohelenalin esters to be relatively high in A. montana buds, but that levels declined as the flowers matured. Schmidt et al.
143 (1998) suggest the presence of a hydrogenase system that converts sesquiterpene lactones from the hydro to the dihydro type and suggest that further investigation into this pathway would be of interest. This study provides more evidence of the existence of such a system.
Considerably fewer sesquiterpene lactones were found in the solvent extract of A. chamissonis compared to those in the literature (e.g. Leven & Willuhn, 1987 and
Willuhn & Leven, 1995). Twenty eight sesquiterpene lactones have been reported in an investigation into the variability of A. chamissonis (Willuhn et al. 1994), while this analysis only found three main components. Analysis of yield plots suggest that the ‘type’ of A. chamissonis corresponds to type E (mostly chamissonolides and arnifolins) identified by Willuhn et al. (1994), and for which the average total content of sesquiterpene lactones reported was 0.62%. The total content in this investigation was approximately 0.17% which, whilst considerably lower than the average 0.7% was above the reported minimum of 0.07% (ibid). In addition, extracting flowers at different stages produced significantly lower levels of total sesquiterpene lactones which suggest that harvesting flowers all at one time could lead to reduced extract quality. The stems and leaves are also very low in sesquiterpene lactone content and if they were present in the final harvest, could further reduce the quality of the extracted material.
However, as mentioned previously, the natural variation of A. chamissonis is high and it may be that by varying the agronomic conditions that levels of sesquiterpene lactones in both A. montana and A. chamissonis can be increased.
144 3.0 Yield of A. montana and A. chamissonis over 6 years
3.1 Introduction
3.1.1 Yield: a balance between quality and quantity
For a ‘conventional’ crop, productivity can be expressed as the rate at which organic matter is accumulated in addition to its own energy requirements.
Presuming no mineral or water deficiency, this is normally a function of both the intensity and duration of the solar radiation (Tivy, 1990) although for a crop such as Arnica this may not reflect the full story. Whilst it is important that the quantity of
Arnica flowers produced is high enough to make the crop financially sustainable, it is also important that the levels of sesquiterpene lactones contained within the
Arnica flower extract are of the right quality and quantity for medicinal purposes.
Whilst the production of a large amount of flower material is important, high levels of the active compounds would increase the quality of the crop and hence its value. The production of secondary compounds is an energy intensive process
(Levin, 1976) and as the amount of carbon available varies throughout the season
(Mooney, 1972), the proportion allocated to defence is likely to depend upon pathogen, pest and other selective pressures.
It is important too, that a balance is struck between vegetative growth and flower production. Heavy application of nitrogen fertiliser for example, can have the effect of encouraging vegetative growth rather than flowers (Forbes & Watson, 1992). As for many crops, only a proportion of the ‘recoverable’ yield can be classed as of commercial value, which in this case is the dried flower material. The proportion of the crop that is utilizable is expressed as the crop (or harvest) index and is the
145 ratio of commercial to recoverable yield (Forbes & Watson, 1992). Hormones such as chlormequat (Tivy, 1990), can be used to create plants with shorter, more uniform stem lengths thereby generating a crop with a higher index, but it is unlikely the application of such treatments would be acceptable to the organic market (Soil Association, 2009).
As defined by Tivy (1990 P21), “in agricultural terms, the crop plant is the basic farm production unit, while the atmosphere and the soil provide the essential resources of carbon dioxide, water, mineral nutrients and solar radiation on which the process of photosynthesis or primary biological production depends”. All plants have different requirements for maximum growth but climate is normally the most limiting factor (Livy, 1990). Hence it is important for any trial that the extent to which environmental factors influence yield in both qualitative and quantitative terms, is determined.
Growing the plant in conditions very similar to its natural environment would perhaps be the most obvious starting point for a study such as this, however field trials of Arnica in similar environments have not yet demonstrated stable agricultural yield (Cassells et al. 1999). In addition, for wild populations of A. montana and A. chamissonis the criteria for fitness are the traits that result in the greatest number of descendents (e.g. seed production and dissemination), whilst in agriculture, yield is of most importance. Even where the seed is the crop it is usually seed number and/or mass that is of top priority, not viability (Jackson &
Koch, 1997). Whilst maximum growth can sometimes be obtained by ensuring growing conditions match those of wild populations, this will not necessarily
146 correspond to high yield just as cloning the most productive plant in the wild, won’t necessarily create a stock ideal for agricultural production.
In addition to striking a balance between growth and yield, it is possible that a balance between weight of flowers produced, and quality of extract will also need to be struck. This is likely to be the case in Arnica if sesquiterpene lactones are involved in plant defence or are produced in response to stress, as found in other species (eg. Guillet et al. 2000 & Saker et al. 2000). For example, it may be that the most favourable conditions for sesquiterpene lactone production is a stress that affects plant growth (Fig 3.1). For such conditions to be identified, field trials are essential.
Frost Rain Wind
Air Weeds pollution
Herbicide Insects & damage other animals
Microbes
Mineral deficiency Wrong pH Poor soil aggregation Poor drainage
Figure 3.1 - Factors that can cause plant stress
Adapted from Forbes & Watson, 1992
To this end, trials monitoring the yield of both A. montana and A. chamissonis were run over a 6 year period, in field sites at the Agronomy Institute (AI) in
147 Kirkwall, Orkney (Section 1.8). Weather data were monitored throughout the trials and soil type and profile was determined, to investigate any correlation between the primary limiting environmental conditions and yield. Together, it was anticipated that this data would contribute towards the aim of identifying key factors for future sites; factors that the sites should contain should these trials be successful and those which might be best avoided should the yield be less than commercially viable. This chapter will describe the results of these field trials, and the subsequent implications for the project as a whole.
3.1.2 The established trials
Seed for the trials was purchased from CN Seeds unless otherwise stated. All of the Arnica trials in this study were planted in 2002 on the same seaward facing slope (Figure 3.2 & 3.3). It was decided to grow Arnica on this site due to its northerly location and long day length during the growing season as well as for its proximity to laboratories equipped for drying the flowers, and for the monitoring that had been ongoing on the site in terms of pesticides, herbicides and weather data. As for any trial, clarity of site history is essential if the number of potential variables which could affect yield is to be limited.
The yield trial areas (AM1, AM2, AC1 and AC2) were established in 2002, prior to the commencement of this study. As Orkney is renowned for its strong winds
(Spence, 1908 and Towrie, 2008) trials were created in which both A. montana
(AM1, Fig 3.4) and A. chamissonis (AC1, Fig 3.5) were sheltered with a screened fence. To determine the effectiveness of this shelter, the remaining two areas of A. montana (AM2, Fig 3.6), and A. chamissonis (AC2, Fig 3.7) were left exposed.
148
Figure 3.2 - Some of the Arnica trial plots
See Figure 3.3 for location
AM3
*
Towards the sea North
Figure 3.3 - Map of the Arnica trials
and mark the location of the two soil pits and * marks the point from which the photo in figure 3.2 was taken
149 The forces applied by winds are normally to the horizontal and proportional to the area of the plant presented to the wind (Forbes & Watson, 1992). In the case of A. montana, the majority of the leaf material is in a rosette close to the ground; hence the flowering stalk of the plant would be the part most susceptible to wind stress, particularly when heavy with flower. However, as for all rosette plants, the flowering head is formed near to the ground and then carried up on a rapidly extending stem for flowering and seeding (Forbes & Watson, 1992), which limits the time that the plant is susceptible to high winds. For A. chamissonis a much higher proportion of the plant material is above ground and so this species may be more subject to ‘lodging’ or structural failure from wind stress.
Figure 3.4 - AM1 (screened)
150
Figure 3.5 - AC1 (screened)
Figure 3.6 - AM2 (not screened)
151
Figure 3.7 - AC2 (not screened)
3.1.3 Harvesting method
All plots were harvested in 2003 prior to the commencement of this study. The resulting data gives an insight into the yield of both species (table 3.1).
From the long harvesting periods described in Table 3.1 and illustrated in Figures
3.8 and 3.9, it can be seen that the flowering for both species is staggered. At first in figure 3.8, the weight of individual flowers is high which is when the apical bud flowers. This is then followed by the blooming of a succession of axillery buds which are smaller in size. This is of advantage to the alpine or northern plant as it means that frost damage to early flowers is not catastrophic to the plant, as later flowers may be more successful (Forbes & Watson, 1992). However, the consequence of staggered flowering in agricultural terms is repeated harvesting and the associated additional costs in terms of time and labour.
152
100 25
80 20
60 15
40 10
Number of flowers/m2 Dry weight of flowers(g/m2)
20 5
0 0
n n n n n n l l l l l l un un un u u u u ul ul u u u u J J J Jun Ju Jun Ju -Jul Jul Ju May - - - 2 4-Ju 6-J 8-J 0-J 8- 0- 2-J 0- 2 4 6 8-Jun 2- 4- 6-J 4- 6- 8-J 1 12-J 14-Jul 16-Jul 1 2 2 24-J 26-Jul 28-Jul 3 10-Jun 1 1 1 18-J 20-Jun 22-Jun 2 2 2 30-J Date Dry weight of flowers Number of flowers
Figure 3.8 - Number of A. montana flowers picked per m2, with dry weight of flowers/m2 for 2003 (AM2)
Line indicates date of 50% plot harvest (weight) Adapted from: Martin, 2003 (unpublished).
250.0 25.0
200.0 20.0
150.0 15.0
100.0 10.0 Number flowers/m2 of
(g/m2) flowers of weight Dry 50.0 5.0
0.0 0.0
l l l l l l l l l 3 3 3 3 3 u u u u u u 0 -0 03 -03 03 03 03 -03 03 0 J J J J J J n- n- n- n- l- 6- 0- un un u u u u Jul-0 u 08-Ju 10- 12-Jul 14- 1 18-Ju 2 22-Ju 24- 26-Jul 28- -Jun- -J -Jun-03 J J J J J 2 4-Jun-036 8 0-Jun- 0-Jun-0 2-Jul-034- 6-J 1 12- 14-Jun-0316- 18-Jun-032 22- 24-Jun-0326- 28-Jun30- Date
No Flowers Flower Dry Weight
Figure 3.9 - Number of A. chamissonis flowers picked per m2, with dry weight of flowers/m2 for 2003 (AC2)
Line indicates date of 50% plot harvest (weight) Adapted from: Martin, 2003 (unpublished).
153
The much lower average flower weight of the A. chamissonis flowers is indicative of their smaller size compared to the flowers of A. montana . Martin (2003, unpublished) concluded that as the flowers of A. chamissonis were smaller and the flowering staggered, the hand harvesting of flowers from this species would not be practicable with an estimated 3.5 hours to pick 1 kg of dried A. montana and 9.2 hours to pick the equivalent for A. chamissonis . With hand harvesting not financially viable, a harvesting trial was conducted in the same year on an A. montana plot. The trial simulated non selective mechanical harvesting with a cutter bar, the plants in each plot being cut at the same height on each occasion (n=7 and 6 plants per plot). The cutting height was varied, but was chosen to make sure that the maximum numbers of open flowers were harvested.
The results of the trial are displayed in Table 3.2 and demonstrate a trend for flower yields to increase with less frequent harvesting. Mechanical harvesting would of course remove flowers at all stages (Fig 3.10), so the less frequent harvesting allows time for buds to develop. The percentage of ‘over-mature’ flowers increases, but as Douglas et al. (2004) found that such flowers produced better quality extract (see also section 2.1.2) it maybe that simulated mechanical harvesting when approximately 50% of apical flowers and 50% of flowering shoots are open, would be the most effective method of harvesting. However, it should be noted that there was little difference between the totals of the harvesting treatments, and that the yield of the hand picked control for this plot was lower than that of the yield trials (AM1 and AM2) outlined previously.
154
Table 3.1 - Summary harvest data for 2003 harvest of A. montana and A. chamissonis in exposed (AM1 & AC1) and sheltered (AM2 & AC2) trials
Trial
AM1 AM2 AC1 AC2 Open (O) or protected (P) P O P O from wind Area harvested (m 2) 4.2 4.2 1.5 1.5 Number of plants in harvested 77 77 18 18 area 30/5 to 30/5 to 30/5 to 30/5 to Harvesting period 28/7 28/7 28/7 26/7 Number of harvests 30 30 30 27
Fresh weight of flowers (g/m 2) 600 564 993 596 Fresh weight of flowers 32.7 30.8 81.4 48.9 (g/plant) Dry weight of flowers (g/m 2) 141 131 208 132
Dry weight of flowers (g/plant) 7.7 7.1 17.1 10.8
Number of flowers per m 2 666 686 2597 1723
Ave. No. of flowers per plant 36.3 37.4 237 152
Average flower weight (g) 0.21 0.19 0.07 0.07
Adapted from: Martin, 2003 (unpublished)
a) b) c)
Figure 3.10 - Flowers of A. montana
a) at bud b) partially open c) over mature
155
Although mechanical harvesting has been shown to be effective, the small harvesting area in this trial (e.g 1.5m 2 for AC1 and AC2) means it would be unlikely to be appropriate or effective in this case and so hand harvesting was adopted throughout.
Table 3.2 - Harvest data from the 2003 harvesting trial of A. montana
Average Flowers at Over mature Flower buds Stem number of correct stage Flowers Treatment times Dry Dry Dry Dry No./ No./ No. / harvested 2 mass 2 mass 2 mass mass m m m per plot (g/m 2) (g/m 2) (g/m 2) (g/m 2)
Hand harvesting 23.7 354 77
Simulated mechanical harvest 3.7 106 27 207 25 15 4 25 when ca 50% open apical flowers
Simulated mechanical harvest 3.4 124 33 189 28 20 6 26 when ca 75% open apical flowers
Simulated mechanical harvesting when ca 3.1 147 35 165 21 33 11 24 50% open apical and ca 50% open flowering shoots LSD n/a 81 19 63 11 15 5 12
Adapted from: Martin, 2003 (unpublished)
3.1.4 The climate
That climate will have an effect on crop yield is undisputed (e.g Thompson & Antill,
1769), and as indicated in Figure 3.1, can be an important source of stress. Any analysis of yield therefore, should investigate potential correlations with both the
156 qualitative and quantitative analysis of the floral extract, and measurements of the environmental factors. To this end, climate data from Hundland Loch Climate
Station (Fig 3.11) in Birsay, Orkney, was utilised. Although this station was slightly further north than the Agronomy Institute (Fig. 3.12), the distance was deemed not to be significant in the terms of this study.
Figure 3.11 - Photos of the climate station at Hundland Loch
Orkney’s northerly latitude results in a summer season of long days, which could lead to the production of high quality plant material. ‘Hours of sun’ is hence an important factor to record. However, as mentioned in section 3.1.2, the Orkney
157 Islands are renowned for one particular aspect of their climate: the wind. In summer force 4 winds are the norm, with an average of force 5 in winter (Table
3.3). For the effectiveness of the screening in AM1 and AC1 to be judged, measurements of wind speed are required. If yield in the sheltered plots is significantly higher in seasons with particularly strong winds, then it may be that screening all Arnica plots would be beneficial.
Figure 3.12 - Map of the climate station at Hundland Loch
The green tab is the climate station and the yellow tab marks the location of the field sites at the Agronomy Institute. Image © Google – Map data © 2009 Tele Atlas PPWK
Whilst there are no clear wet and dry seasons in Orkney: the days of rain are fairly constant, whilst the monthly rainfall is quite variable (Table 3.3). Total and maximum rainfall will hence also be investigated, in case waterlogging is associated with poor yield.
158
Table 3.3 - Kirkwall monthly averages (1971-2000)
Days of Days of Wind Wind Max Min Air Sun Rain Rain at at Temp Temp Frost >= 1mm 10 m 10 m [deg Month [deg C] [days] [hours] [mm] [days] [knots] [force] C] Jan 6.1 1.7 6.9 28.8 109.8 20.0 16.8 5
Feb 6.2 1.6 6.0 58.5 85.7 15.9 15.8 5
Mar 7.3 2.2 5.0 94.6 92.8 17.1 15.6 5
Apr 9.0 3.3 2.8 128.4 59.4 12.7 13.3 4
May 11.7 5.4 0.7 181.0 49.4 10.0 12.0 4
Jun 13.7 7.7 0.0 150.0 53.7 10.3 11.3 4
Jul 15.5 9.8 0.0 129.0 55.1 11.2 10.8 4
Aug 15.7 10.0 0.0 133.9 65.9 12.0 10.7 4
Sep 13.6 8.4 0.0 102.9 99.3 16.7 12.9 4
Oct 11.2 6.5 0.2 73.8 113.6 18.4 14.5 5
Nov 8.3 3.8 2.3 38.1 129.8 20.5 15.2 5
Dec 6.7 2.4 5.0 20.8 115.1 19.6 15.6 5
Average 10.5 5.3 28.9 1139.7 1029.4 184.4 13.6 4
From Metoffice, 2009
Although frost stress is included in figure 3.1, extreme frosts during flowering are unlikely as the climate in Orkney is free from the extremes of continental Europe.
This could still impact on yield though, if winter vernalization (a phase during which plants need a minimum temperature of 2-8°C) is a re quirement for flowering.
Although no evidence of this requirement has yet been found in the literature, as
Arnica is an alpine plant (Pegtel, 1994) it is likely to be the case and so winter minimum temperatures should be monitored.
159
3.1.5 The soil
Soil is central to effective land use (Ellis & Mellor, 1995), so it is important when investigating the properties of any crop that the nature of the substance in which it is growing is defined. Soil is composed of organic matter, mineral material, the roots of plants, microbial and animal biomass (Killham, 1994), all of which contribute to its structure and texture. The varying ratios of clay, silt and sand determine the texture of the soil (Ellis & Mellor, 1995) (Fig 3.13). Light soils (over
80% sand) are coarse textured and normally described as dry due to the rate which water percolates through them. Heavy soils are more finely textured (over
25% clay) and are both water and nutrient retentive.
The particles within the soil can join together via electrostatic forces to form
‘domains’ which can be up to 5 µm in diameter. Larger ‘peds’ or ‘aggregates’ of up to 250 µm can be formed via fungal hyphae and plant roots and the gaps between these are known as pores or voids (Ellis and Mellor, 1995). Within the pores can be found the gaseous component (a mixture of gases derived from the atmosphere), and the water component or soil solution. The structural aspects of soils are very important as the rate to which a root can extend into the soil will be influenced by the degree to which it must exert pressure to enlarge soil pores.
Pore space decreases with clay content and with soil depth. Decreased pore volume provides both mechanical impedance as well as influencing water regime, nutrient flow and aeration (Killham, 1994). Sandy and silty soils do not aggregate easily; whilst those with a significant proportion of clay are sticky, shrink when dry and expand when wet (Tivy, 1990).
160
Figure 3.13 - Chart showing the percentages of sand, silt and clay and the textural classifications
From: USDA Natural Resources Conservation Service , 2007
Soil survey maps describe the area of Orkney in which the agronomy institute is located as Tresdale and Bilbster (Table 3.4), which are both poorly drained drifts
(or superficial deposits as they are also known), derived from old red sandstone.
However, it is important in any yield study to characterise the immediate area under investigation and to this end soil profiles are normally constructed (Ellis &
Mellor, 1995 and see section 3.2.1).
161
Table 3.4 - Soil type
Soil type Description
Tresdale Poorly drained drifts* derived from sandstone, flagstone, (marls), mudstones and limestones of the Stromness and Ronsay (and the Eday) beds of the middle old red sandstone
Bilbster Poorly drained drifts derived from sandstone, flagstone, marls, mudstones and limestones of the Stromness and Ronsay and the Eday beds of the middle old red sandstone
Source: Soil Survey of Scotland, Fulton, 1979.
Soils contain the nutrient pool from which the plant draws both the macro and micro-nutrients essential for growth. The degree to which a nutrient is retained depends upon the soil’s texture and structure as well as its cation exchange capacity (Tivy, 1990). Negatively charged clay minerals in the soil tend to attract nutrient cations (e.g. Ca 2+ ), although they are readily removed back into solution by ion exchange. Nutrients not attracted by ion exchange such as nitrate, tend to be leached more readily (Killham, 1994). In soils that have a neutral pH, the exchange sites tend to by filled by ions such as calcium (Ca2+ ) (ibid). A lack of available minerals can lead to plant stress and so for the purposes of this study, an analysis of soil mineral composition, aggregation and pH was conducted.
3.2 Materials and Methods
3.2.1 The site
All the Arnica plots in this study were planted on the same seaward facing slope
(Figure 3.1) at the Agronomy Institute of Orkney College UHI (Section 1.8). This
162 was in part due to the availability of the site as well as the proximity to laboratories for flower drying facilities.
Prior to the commencement of this study, seeds were added to a light, peat free compost and kept moist and in a polythene tunnel until germination. Trays were then transferred outside for hardening, then planted into the yield plots at a plant density of 183,333/ha on Friday 17 th May 2002. A moderate amount of fertiliser was added to all four plots before the plot was created and then in April at the start of each subsequent field season. Nitrogen was applied as ammonium nitrate, phosphorous as phosphate and potash as potassium oxide, combined together in the form of slow release solid pellets (N:P:K, 9:19:15). Application rates were equivalent to 40 kg/ha of N, 90.0 kg/ha of P 2O5, 67.4 kg/ha of K 2O.
Although the overall soil types have been described above, this gives little indication as to the different horizons present and their depth, which is important if the soil in the area is to be fully characterised. To this end, two soil profile pits were dug to a depth of 1 metre in March 2007 and soil samples from the A 1 horizon were analysed independently by the Scottish Agricultural College (SAC) for pH and available minerals.
3.2.2 Meteorological data
The equipment in the climate station described in section 3.1.4 (Figs 3.11 & 3.12) consisted of mercury thermometers (wet, maximum, minimum and grass), all housed in a Stevenson Screen and supplied by the Met Office. The rain gauge Mk
2 and the tapered rain measure were both supplied by the Met Office, as was the
163 Campbell Stokes Universal sun recorder Mk3C and the sun measurement cards.
The mercury barometer was a Fortin pattern and the precision aneroid barometer was produced by Negretti & Zambra and checked by the Met Office.
The anemometer and wind vane were produced by Davis Vantage Pro2 and recordings were monitored by a Davis data logger with a Weatherlink to PC. All readings except sun hours were taken at 0900 UTC and the sun hours card was changed each day after sunset.
3.2.3 Flower harvesting
Flowers were harvested between 14:00 and 16:00 hours after the methodology of
Galambosi et.al (1998). During these hours the flowers were judged to be driest and hence would take the shortest time to dry till constant weight. Flowers were cut just below the flower, taking care to minimise the amount of stem present.
Flowers from each plot were weighed and dried at 40°C till constant weight
(normally 4 days). After this, flowers from each plot were stored in labelled brown paper bags at c.a. 4°C till later analysis in Inver ness.
Water constitutes a large proportion of the weight of fresh Arnica flowers and so small variations in the moisture content could have significant yield effects. For such variation to be excluded, flower yield will mainly be expressed as dry weight per square metre. Yield is more usefully calculated on this basis, rather than per plant because it can be very difficult to tell where one plant stops and the next begins (personal observation).
164 Initially, all flowers were harvested from all of the plots, but this soon became impracticable for AC1 and AC2 due to the number of flowers involved. Instead, a
1.5m 2 sample section was marked out within each plot and this was harvested instead.
3.2.4 Extraction
The extraction of the dried flowers was performed as outlined in section 2.2.2.2.
Santonin was used as an internal standard and was used both as a reference compound and to calculate recovery (Wagner& Merfort, 2007). All tinctures prepared were then stored below 4°C at ca -20°C pri or to analysis.
3.2.5 Analysis
The material was analysed both quantitatively and qualitatively as described in section 2.2.3.
3.2.6 Statistical analysis
As with most yield trials, the need to strike a balance between running trials on new crops and taking land away from established ‘guaranteed yield’ crops, means that limited land is available in the early stages of experimentation. As a result, the resources available are often less than desired and often, early ‘look and see’ stages, of which these yield plots are representative, do not involve enough replication for statistical tests to be meaningful (Mead et al, 1993). However, some basic correlation was conducted (Genstat) to determine whether any trends between environmental factors and yield were evident, in order to determine the ideal conditions for future trials.
165 3.3 Results
3.3.1 The soil
The two soil pits (Figs 3.14 & 3.15) were both dug to a depth of 1 metre to the sandstone bedrock. The characteristics of each profile were assessed and since the horizons in each pit were essentially the same, with minor variation in depth, the characteristics for both are summarised in Table 3.5. No difference was found between the horizons of the two pits, although the depths of each varied slightly.
Table 3.5 - Soil profile characteristics
Colour & Horizon Texture* Structure Consistency Roots Comments mottling
Dark grey Blocky/crumb A Silty clay Friable Few 1 brown Strong structure
Blocky/crumb A Grey brown Silty clay Friable None 2 Strong structure
Grey brown Blocky/crumb B with rusty Clay loam Firm None Strong mottling structure Mottled Brick With some red with Prismatic/ induration orange, platey (hardening, C Clay Very Firm None yellow light Strong comes with grey and dark structure lots of grey rain).
*Texture after Landon, 1988
166
Soil Pit 1
0cm
Dark grey brown – Horizon A 1 8cm
Grey brown – Horizon A 2 21cm Grey orange - Horizon B 31cm Mottled orange, yellow, red
and brown - Horizon C 80cm Bedrock –Horizon D 90cm
Soil Pit 2
0cm
Dark grey brown - Horizon A 1
13cm
Grey brown – Horizon A 2
30cm Grey orange – Horizon B 39cm
Mottled, orange, yellow, red and grey – Horizon C 79cm
90cm Bedrock – Horizon D
Figure 3.14 - Diagram of both soil pits illustrating the similar nature of the two pits
167
A1
A2
B
C and D
Figure 3.15 - Detailed photos of soil pit 1 illustrating different horizons.
As the roots of the Arnica plants were predominantly in the A horizon, they were in the silty clay area which is composed of roughly 50-60% silt, 40-50% clay and 0-
10% sand. The grey colour of Horizon A could be the result of leaching due to high rainfall; this is also indicated by the presence of induration in Horizon B. However, the rusty mottled colour caused by the presence of hydrated oxides of iron, in horizon C indicates that drainage is often impeded at the site, despite the seaward facing slope. Poor drainage is likely to cause A. montana difficulties as it has been reported that it is more likely to be found on well-drained podzols (Pegtel, 1994), while A. chamissonis is often found in wet places (Rickett, 1973). However,
168 although the rainfall in Orkney is high, the soil is on a seaward facing slope and this may lead to better drainage than might normally be expected for a clay soil.
The acidity of the soil was pH 6.0 which is at the alkali end of the range recommended by Pegtel (1994), but still within acceptable limits for A. montana. A. chamissonis, on the other hand, is more commonly found on more alkaline soil
(Rickett, 1973).
The extractable mineral content was extracted independently (see section 3.2.1) and the results are presented in Table 3.6. The levels of phosphorous, potassium and magnesium were deemed to be moderate. Although there is no official average set by the SAC for calcium, the levels recorded were deemed no higher than the ‘average’ soil that they use for comparative purposes. Low calcium is important as Raison et al. (2000) and Jenelten & Feller (1992) state that for A. montana , calcium levels should be low - less than 1% (10,000mg/L). The results
(Table 3.6) show that the soil at the Agronomy Institute is well below this level.
Table 3.6 - Extractable mineral content
Extractable mineral mg/l Phosphorous 10.2 Potassium 118 Magnesium 125 Calcium 1110 Aluminium 16.3
SAC Advisory soil report, 2006
169 3.3.2 Climate
Climate data from Hundland Loch Climate station are summarised in Table 3.7.
Data have been blocked into months influencing the summer field seasons (i.e. running from August to the following July), so that potential correlations between yield and environmental conditions can be investigated. The first trial was planted out in May 2002, and whilst A. montana did not flower till the following year, A. chamissonis flowers were harvested during August 2002. As a result, the applicable 2002 weather data runs only from May to July and so cannot easily be compared with that of subsequent years. For clarity, the 2002 recordings are shown in the table, but are not presented in Figures 3.16-3.17.
Of course, some environmental factors such as wind speed may be more influential on Arnica in the summer months (May-July) when the plant is flowering rather than in the winter months when the plant is below ground. In order to investigate this further, climate data just for the summer months were compiled and are summarised in table 3.8. As can be seen from the table, variation was less over this shorter time period, with average wind speed, number of gale days, average temperature and minimum temperature all remaining fairly constant. The highest rainfall in one day and highest total rainfall was in the summer of 2007 however, which is also when the lowest temperature, lowest maximum temperature and highest wind speed was recorded for the months in question.
170 Table 3.7 - Orkney climate data
Year 2002 2003 2004 2005 2006 2007
Factor May-Jul Aug-Jul Aug-Jul Aug-Jul Aug-Jul Aug-Jul
Mean Min Temp (°C) * 8.33 6.0 5.6 5.9 5.5 6
Mean Max Temp (°C)* 15.2 11.9 11.5 11.4 11.4 11.7 Mean of maximum and minimum temperature 11.7 9 8.6 8.6 8.5 8.9 (°C) Min Temp (°C) (lowest -3.9 -6.4 -8 -5 -5 -3.2 daily temp recorded) Max Temp (°C) (highest daily temp 20.8 24.5 26.5 22 24.6 24.5 recorded) Min Grass Temp (°C) -3.8 -10.4 -13 -9 -9.5 -4 Av Rel Hum at 0900 X X X 81 81.8 83.2 (%) Total rainfall for year 174.3 828.5 890.7 1029.7 905.3 1263.1 (mm) Maximum rainfall in a 28.7 21.2 20 29.1 31.3 45.8 day (mm) Days with Rain X X X X 269 176 Days with No Rain X X X X 96 189 Total Global Radiation 2 X X X X X 2799726 (kJ/m )
Total sun (hours) 609.7 1304.6 1245.4 1151.8 1257.9 1176.4 Maximum sun in a day 15.5 15.7 14.7 15.1 15.1 14.6 (hours) Average Wind Speed 10.07 10.7 11.1 12.6 11.2 12.2 (Knots) Wind Direction SE SE SW S S S Wind Maximum 42.6 64.3 67.8 82.6 65.2 69.6 (Knots) Number of gale days 1 9 13 14 11 11 Minimum wind chill -5.7 -12.7 -9.9 -7.9 -8.6 -8 (°C) Air Pressure minimum 980.9 971.2 969.3 961.9 969.6 958 (mb) Air Pressure maximum 1031.7 1040.5 1038.3 1038.4 1042.7 1038.6 (mb)
‘X’ represent years for which data are not available. The grey section was compiled from information from the Metoffice (2007), all other weather data was generously provided by Keith Johnson at the Hundland Loch Climate station. Although there was some overlap in the data available from both sources, correspondence was within =/-0.5 of a degree. For the purposes of data comparison, table 3.7 is compiled mostly from the more detailed information available from the Hundland Loch Climate station. *Means were calculated by adding all the daily maximum or minimum temperatures and dividing by the number of days in the month. Minimum wind chill represents the lowest recorded for the year
171 Table 3.8 - Climate data for the May-July period of each field season
Climate variable (May-Jul) 2003 2004 2005 2006 2007
No. of Gale days 0 0 0 0 1
Average wind speed (knots) 9.5 10.2 10.2 9.7 10.3
Maximum wind speed (knots) 43.5 36.5 41.7 47.8 48.7
Wind chill (°C) -4.7 -0.2 -1.5 -0.8 -2.9 Total Rainfall for May – July 178.4 159.2 155.7 120.3 237 (mm) Maximum daily rainfall (mm) 16.2 13.6 18.9 13.2 42 Maximum daily temperature 24.5 21 21.4 24.6 19.4 (°C) Minimum daily temperature (°C) -0.4 1 0 0.6 -0.9
Minimum daily grass temp (°C) -3 -2.3 -2.7 -2.6 -3
Average daily temperature (°C) 12.6 11.5 11.1 12.7 10.7
Total sun (hours) 499.3 430 415.1 554.3 452.6
The temperature data displayed in Figure 3.16, indicate that the mean temperatures for 2003 – 2007 generally were fairly constant although in 2004 both the highest (26.5°C) and lowest (-8°C) temperatures were recorded, as well as the lowest grass temperature (-13°C). The mildest year was 2007, with a mean temp of 8.9°C, a maximum of 24°C and a minimum of -3.2°C . The climate data in Table
3.8 is shown graphically in Figs 3.16-3.19 in order to illustrate more satisfactorily the year-to-year differences in variables.
172
30.0
25.0
20.0
15.0
10.0
5.0
0.0 2003 2004 2005 2006 2007 Temperature(degrees C) -5.0
-10.0
-15.0 Year
Mean Min Temp Mean Max Temp Min Temp Max Temp Min Grass Temp Wind Chill
Figure 3.16 - Graph of temperature data from 2003-2007
Figure 3.17 reveals an increase in maximum single day rainfall between 2003 and
2007, alongside a general increase in total rainfall. In contrast, Figure 3.18 indicates a general downwards trend in sun levels, with the highest total and maximum sun recorded in 2003.
The maximum wind, average wind speed and number of gale days (Fig 3.19) were fairly constant. The maximum wind recorded was normally around 65 knots, the average was about 11 knots and there were on average 12 gale days each year.
All of these measurements were highest in 2005 however, when there were 14 gale days, an average wind speed of 12.6 knots and the maximum wind recorded was 82.6 knots.
173
250
200
Level (mm) 150
100 2003 2004 2005 50 2006 2007 0 2006 2007 Total 2005 Maximum 2003 2004 Rainfall rainfall
Figure 3.17 – Graph of rainfall recordings from 2003-2007
1350.0 15.8
15.6 1300.0 15.4
1250.0 15.2
15.0 1200.0 14.8
1150.0 14.6 Sunshine (hours)
Total Total sunshine (hours) 14.4 1100.0 14.2
1050.0 14.0 2003 2004 2005 2006 2007 Year
Total sun (hrs) Maximum sun (hrs)
Figure 3.18 - Graph of recorded sun from 2003-2007
174
90 16
80 14 70 12 60 10 50 8 40 6
Wind Max (kn) 30
4 Number of gale days 20
10 2 0 0 2002 2003 2004 2005 2006 2007 Year Wind Max No. Gale days
Figure 3.19 - Graph of maximum recorded wind speed and gale days from 2003- 2007
3.3.3 Yield
A variety of yield data were collected and are presented for both A. montana
(Table 3.9) and for A. chamissonis (Table 3.10). The average flower weight was quite consistent for both species, and as yield is the product of the number of flowers and the average flower weight, changes in yield are largely the result of changes in flower or plant number.
For A. montana the highest yield (140.7g/m 2) was obtained in 2003 in AM1, the sheltered plot. This equates to 1407 kg/ha, much more than the 1000 kg/ha reported both in Germany (Bomme and Daniel, 1994) and in Finland (Galambosi et al. 1998). The following year however, the yield in AM1 dropped sharply
175 (71g/m 2) to the extent that it was abandoned in 2005 due to the low numbers of surviving plants (7). The yield of AM2 also dropped sharply and although this decline was not as severe as for AM1, the total number of flowers, number of flowers per plant, number of surviving plants, fresh weight and dry weight per hectare all declined year on year till 2007, when a dried flower yield of 10.5 g/m 2 was produced by the 40 remaining plants. Overall, an average of 55.08 g/m 2 was produced each year. This translates to 551 kg of dried flower material produced per hectare which can be favourably compared to yields of 94 to 284kg per hectare for A. montana in Italy (Bezzi & Ghidini, 1989), but is much less than the
German and Finnish trials.
The cause of this decline in yield was a decrease in plant number which is thought to have been caused by disease. The affected A. montana plants first displayed yellow leaves, then stem, base and crown rot typical of an oomycete infection.
Two sample plants were sent to DEFRA for rapid serological testing by lateral flow device (LFD) which detected the plant pathogens Phytophthora and Pythium , both of which are favoured by waterlogging (DEFRA, 2005). Although a number of plants in AM2 displayed similar symptoms, the disease did not spread as rapidly in this plot. It may be that environmental factors such as high rainfall have contributed to the spread of this disease, or it may be that agronomic conditions have exacerbated this condition.
176 plants surviving of Number dryweight flower Average of fresh as % Dry weight m2 / (g) Dry weight (g) dry weight Total (g) weight Total fresh m2 / flowers of Number of flowers number Total Treatment Year Grey sections represent foryears which data is not AM1 77
2002
AM2 77 77
2519.8 2519.8 140.7 590.7 665.7 available. In 2002 the plants had not yet flowered 2796 2796 AM1 23.4 0.2 72 72 2003
Table 3.9 - Yield data for AM1 and data3.9AM2 forTableYield AM1 - 2368.4 2368.4 130.9 130.9 549.7 685.7 2880 AM2 23.2 0.2 72
298.3 350.2 1532 1471 AM1 19.5 0.2 34 34 71 2004
383.6 383.6 478.6 2051 2010 AM2 18.7 91.3 0.2 69
AM1 inand 2005 AM1 was abandoned todue disease. 7
2005
113.6 150.5 AM2 16.4 691 632 0.2 68 68 27
AM1 0 0
2006
338.5 AM2 19.5 15.7 65.8 91.2 383 0.2 58 58
AM1 0 0
2007 231.5 AM2 10.5 43.9 57.9 243 0.2 19 40
177 Average flower dry weight dryweight flower Average of fresh as % Dry weight m2 / (g) Dry weight (g) dry weight Total (g) weight Total fresh m2 / flowers of Number of flowers number Total Treatment Year 950.11 950.11 2563.3 141.8 141.8 212.7 3845 AC1 0.06 22.4
2002 2002 1096.82 1096.82 2029.3 2029.3 104.7 104.7 157.1 3044 AC2 0.05 14.3
1489.5 1489.5 2887.3 208.5 312.8 4331 AC1 0.07 21 21
2003 1854.7 1854.7 197.9 197.9 894.6 2782 AC2 0.07 22.1 132 132
1007.3 1007.3 535.7 1511 AC1 0.07 20.7 Table 3.10 - Yield data for AC13.10 dataand- forTable Yield AC2 111 74 74
2004 101.6 101.6 961.3 1442 AC2 0.07 21.3 67.8 477 477
339.84 339.84 608.7 AC1 0.07 19.3 43.6 65.4 913
2005 477.98 477.98 861.3 861.3 1292 AC2 0.07 18.9 60.3 90.4
611.09 611.09 130.9 1368 2052 AC1 0.06 21.4 87.3
2006
504.47 504.47 1062.7 111.1 1594 1594 AC2 0.07 74.1 22 22
238.66 238.66 AC1 0.06 20.3 32.3 48.4 518 777
2007 218.9 488.7 AC2 0.06 21.1 30.8 46.2 733
178 The peak yield for A. chamissonis was also in the sheltered plot (AC1) in 2003
(141.8 g/m 2 or 1418 kg/ha). In all years other than 2005, AC1 produced a greater dry weight per hectare than AC2, the difference varying considerably between years. Although the yield of A. chamissonis also generally declined after 2003, the decline was not as severe as for A. montana and there was an increase in 2006 to
87.3 g/m 2 for AC1 and 74.1 g/m 2 in AC2.
3.3.4 Competition and herbivory
It was observed that A. chamissonis produced many new shoots from its roots each year (Fig 3.20), all slightly away from the last shoot. The rosette leaf arrangement that it adopts at first, allows it to take better advantage of available space at the start of the year when the land is clear (Fig 3.21). Then, when flowering, it bolts and becomes higher than any of the weeds around it. This growth habit allows it to over-shade and hence out-compete weeds more effectively than A. montana which puts up shoots immediately next to the ‘mother’ plant (Fig 3.22), leaving a greater proportion of open area to be colonised by weeds (Fig 3.23). These include dandelions and daisies (which also have the rosette growth habit), and grasses which very quickly become taller than the A. montana . These observations indicate that t is likely A. montana will face more competitive stress in Orkney than A. chamissonis.
179 1cm
Figure 3.20 - Photo of A. chamissonis shoots
John Wishart, 2008
10cm
Figure 3.21 - Photo illustrating the even cover of A. chamissonis plants.
John Wishart, 2008
180 1cm
Figure 3.22 - Photo illustrating the clumped nature of A. montana shoots
John Wishart, 2008
10cm
Figure 3.23 - Photo illustrating the rosette form of A. montana plants.
John Wishart, 2008
181
The two phytophagous insects specific to A. montana : Tephritis arnicae L.
(Diptera, Tephritidae) and Melanagromyza arnicarum Her . (Diptera, Agromyzidae)
(Scaltriti, 1985), are not native to Orkney and were not found on the plants. Slug damage was more common, but was minimal on these plots (personal observation), and was certainly not to the level that would be deleterious to the plant (Fig 3.24 & 3.25). Scheidel & Bruelheide (1999) found the leaves of A. montana to be highly palatable to slugs, but that predamaged leaves were less palatable to slugs than undamaged leaves. This suggests that there is a mechanism by which the production of defensive compounds may be induced, which may deter the slug from ’over-indulging’. Slugs have been noted to eat the flower rays however, and leave the rest of the flower intact as can be seen in
Figure 3.25. Also in that figure is the ‘cuckoo spit’ of Philaenus spumarius L. which although widespread, is not thought to be damaging to the plants.
Figure 3.24 - Slug damage on the leaves of A. chamissonis
John Wishart, 2008
182
Figure 3.25 - Slug eating the leaves of A. chamissonis
John Wishart, 2008
Figure 3.26- Slug damage to flowers of A. chamissonis, with cuckoo spit
John Wishart, 2008
183 3.3.5 Correlations between yield and environmental conditions
It is important to note that for all of the environmental factors, even if a correlation between an environmental factor and yield is found, no causative link can be proved. It may be for example; that the environmental factor in question is correlated with another factor which has more influence, or that the limited sample size has led to a false positive result. As stated previously, there are not enough repetitions for AC1, AC2 and AM2 (n=5, 2003-2007), for any relationships to be stated with confidence, whilst AM1 has only two years worth of yield data and so any correlations found would be meaningless.
The average flower weight of Arnica is fairly constant (section 3.3.3), and so variation in yield will normally be the result of an increase in flower number. It could therefore be predicted that any environmental factor that correlates with yield expressed as flower dry weight/m 2, would also be correlated with flower number. To confirm whether this was the case, both factors were investigated for all environmental conditions, although the results are not all presented in order to limit repetition.
Scatter plots were used to investigate the extent to which the main climate factors correlate with yield (Figs 3.27-3.40) and the Pearson Correlation Coefficient (a measure of the strength between the two variables) was employed to test the significance of any possible relationships (Table 3.11).
184 Table 3.11 - Table of the correlation coefficients for the main climate factors, number of flowers and dry weight, for each plot
Climatic variable AC1 AC2 AM1 AM2
Dry Number Dry Number Dry Number weight of weight of weight of (g/m 2) flowers (g/m 2) flowers (g/m 2) flowers
Wind chill (°C) -0.952 -0.933 -0.912 -0.915 -0.948 -0.886
Summer wind chill -0.637 -0.630 -0.509 -0.527 -0.417 -0.415 (°C)
Minimum temp. (°C) -0.444 -0.408 -0.540 -0.523 -0.771 -0.850
Minimum grass
-0.422 -0.398 -0.564 -0.545 re of number small to due possible not Correlations -0.649 0.724 temp. (°C)
Maximum temp (°C) 0.186 0.186 0.095 0.092 0.406 0.400
Mean max temp 0.674 0.659 0.510 0.526 0.609 0.503 (°C) Summer max temp 0.751 0.784 0.807 0.808 0.376 0.296 (°C)
Mean temp (°C) 0.506 0.488 0.338 0.356 0.452 0.347
Summer mean 0.765 0.800 0.794 0.796 0.434 0.363 temp (°C)
Wind Max (knots) -0.571 -0.602 -0.433 -0.445 -0.417 -0.353
Average wind -0.799 -0.806 -0.800 -0.798 -0.749 -0.724 speed (knots) Average summer
-0.881 -0.910 -0.875 -0.882 petitions. -0.498 -0.389 wind speed (knots) Number of gale -0.740 -0.769 -0.578 -0.598 -0.393 -0.252 days
Total rainfall (mm) -0.720 -0.716 -0.830 -0.818 -0.701 -0.705
Maximum daily -0.604 -0.572 -0.734 -0.717 -0.797 -0.843 rainfall (mm)
Hours of sun 0.869 0.803 0.827 0.830 0.729 0.670
Summer sun hours 0.466 0.529 0.413 0.425 -0.019 -0.121
Maximum number 0.870 0.872 0.915 0.917 0.576 0.476 of sun hours
Correlation (r crit =0.8783, p<0.05) Not significant but above 0.8 Not significant
185
For both A. montana and A. chamissonis , a relationship between flower dry weight and increasing annual wind chill was immediately apparent for AC1 (r (3) =-.952, p<.05), AC2 (r (3) =-.912, p<.05), and AM2 (r (3) =-.948, p<.05) with colder wind chill being associated with both greater dry weight production (Fig. 3.27) and a higher number of flowers produced (Fig 3.28). When just the summer data were analysed, no such correlation was found (Table 3.11) and it should be noted in both Arnica species there is no above ground plant material in winter and so the correlation is likely to be due to a correlation between annual wind chill and another environmental factor. For AM2 a similar, if not so significant trend was also detected for minimum temperature (Fig 3.30). It could be that this is evidence of low temperature stimulating flowering by vernalization. However, the correlation with low temperature was weaker than for annual wind chill and so it is possible the result is due to the association of the latter with another as yet unknown factor, such as damage to surrounding weeds.
186
a) Dry w eight (g/m2) -4 0 50 100 150 200 250 300 350 -5 -6 -7
-8
-9
-10 -11
Minimumwind chill(oC) -12
Minimum annualwind chill (°C) -13
-14
♦ AC1 AC2
b) Dry weight (g/m2)
-4 0 20 40 60 80 100 120 140 160 -5
-6
-7
-8
-9
-10
-11
Minimumwind chill(oC) -12 Minimum annual wind chill (°C) Minimumannual wind chill (°C) -13
-14
♦ AM1 AM2
Figure 3.27 - Scatter plot of minimum annual wind chill against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
187
a) Flow er number -4
-5 0 1000 2000 3000 4000 5000 -6 -7
-8 -9 -10
-11
Minimum wind(oC) chill -12 -13 Minimum annual wind chill (°C) Minimumannual wind chill (°C) -14
♦ AC1 AC2
b) Flower number -4 0 500 1000 1500 2000 2500 3000 3500 -5
-6
-7
-8
-9
-10 -11
Minimumwind (oC) chill -12
-13 Minimum annual wind chill (°C) Minimum annualwind chill (°C) -14
♦ AM1 AM2
Figure 3.28 - Scatter plot of wind chill against flower number for a) AC1 and AC2 and b) AM1 and AM2
188
a) Dry w eight (g/m2) -2 0 50 100 150 200 250 300 350 -3
-4
-5 -6
-7
-8
Minimum temperature (oC) -9
-10
♦ AC1 AC2
b) Dry w eight (g/m2)
-2 0 20 40 60 80 100 120 140 160 -3
-4
-5
-6
-7
-8
Minimum temperature (oC) Minimumtemperature -9
-10
♦ AM1 AM2
Figure 3.29 – Scatter plot of minimum temperature against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
189
a) Flower number -2 0 1000 2000 3000 4000 5000 -3
-4
-5 -6
-7
-8
Minimum temperature (oC) -9
-10
♦ AC1 AC2
b) Flower number -2 0 500 1000 1500 2000 2500 3000 3500 -3
-4
-5
-6
-7
-8 Minimumtemperature (oC) -9
-10
♦ AM1 AM2
Figure 3.30 – Scatter plot of minimum temperature against flower number for a) AC1 and AC2 and b) AM1 and AM2
190 There appears to be a link between average wind speed and the yield of A. chamissonis (Fig 3.31), with greater wind speeds linked with lower yield. However, no such relationship was evident for A. montana and the relationship was not significant in A. chamissonis for number of gale days (Fig 3.32) and no relationship between summer gale days or maximum summer wind speed and yield was apparent, which suggests that lodging or the breaking of stems is not a major factor. However there did seem to be a relationship between average summer wind speed and both the dry weight and flower numbers produced by A. chamissonis (Fig 3.33). The correlation was significant for both AC1 and AC2 (with the exception of the dry weight of AC2), including the sheltered A. chamissonis plot, not perhaps a plot that would be expected to suffer as much wind related stem damage. This could indicate that the shelter was ineffective, that wind is associated with another environmental factor or that wind stress influences flowering perhaps by the diversion of resources away from flowers to the production of stronger roots or stems. While there was no significant correlation
(r=0.603) between total rainfall and average wind speed for the summer months
(Fig. 3.34), there was an indication of a relationship between average summer wind speed and total sun hours (r=-0.793) (Fig 3.35), with increased average wind speed linked to less total hours of sun.
191
a) 13
12.5 12 11.5 11
10.5 10 9.5
9
Average wind speed (Knots) 8.5 8 0 50 100 150 200 250 300 350 Dry weight (g/m2)
♦ AC1 AC2
b) 13 12.5
12
11.5
11
10.5
10
9.5
9 Averagewind speed (Knots) 8.5 8 0 20 40 60 80 100 120 140 160
Dry weight (g/m2)
♦ AM1 AM2
Figure 3.31 - Scatter plot of average wind speed against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
192
a)
15
14
13
12
11
10 Numberof gale days 9
8 0 50 100 150 200 250 300 350
Dry w eight (g/m2)
♦ AC1 AC2
b) 15
14
13
12
11
10
Number ofgale days
9
8 0 20 40 60 80 100 120 140 160
Dry weight (g/m2)
♦ AM1 AM2
Figure 3.32 - Scatter plot of number of gale days against dry weight for a) AC1 and AC2 and b) AM1 and AM2
193 a) 10.4
10.2
10
9.8
9.6
9.4
9.2 Average summer Average wind speed (Knots)
9 0 50 100 150 200 250 300 350
Dry w eight (g/m2)
♦ AC1 AC2
b) 10.4
10.2
10
9.8
9.6
9.4
9.2 Average summer wind summer Average speed(Knots) 9 0 20 40 60 80 100 120 140 160
Dry w eight (g/m2)
♦ AM1 AM2
Figure 3.33 - Scatter plot of average summer wind speed against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
194 1400
1200
1000
800
600 Total rainfall (mm) rainfall Total 400
200
0 9 9.5 10 10.5 11 11.5 12 12.5 13
Average windspeed (Knots)
Figure 3.34 - Scatter plot of total rainfall against average wind speed
600
550
500
450
400
350 Total summer sun hours sun Totalsummer 300
250
200 9.40 9.50 9.60 9.70 9.80 9.90 10.00 10.10 10.20 10.30 10.40 Average summer windspeed (Knots)
Figure 3.35 - Scatter plot of total summer sun hours against average summer wind speed
195 The relationship between total hours of sun and both flower dry weight and flower number was investigated next, and the scatter plot for dry weight is presented in
Figure 3.36. A relationship by which increased hours of sun appeared to correspond to increased yield was evident, with a similar pattern found for maximum sun hours (Fig. 3.37) for both flower number and flower dry weight.
Although all relationships looked strong, only AC2 had a significant correlation with maximum sun hours and when the analysis was run again, this time just using the total hours of summer sun, the relationship was less strong (Fig 3.38). This could indicate that the hours of sun that occur after flowering are important for Arnica to build up rhizome reserves for the following year.
As before, any apparent relationship may be the result of hours of sun being correlated with another factor, such as temperature. However, the total hours of sun are not correlated with maximum temperature (r=0.572) and are not correlated with mean temperature (r=0.234) but are strongly correlated when just the summer months are taken into account (r=0.831) (Fig 3.36). No relationship was found between maximum temperature or mean temperature and yield, but there did seem to be indications of a relationship between summer mean temp and both flower dry weight (Fig 3.37) and flower number (Fig. 3.38) for A. chamissonis , although this was not significant.
196
a) 1350
1300
1250
1200
1150
1100 Total hours of sun
1050 1000 0 50 100 150 200 250 300 350
Dry w eight (g/m2)
♦ AC1 AC2
b) 1350
1300
1250
1200
1150
1100 Total hoursof sun
1050
1000 0 20 40 60 80 100 120 140 160
Dry w eight (g/m2)
♦ AM1 AM2
Figure 3.36 - Scatter plot of total hours of sun against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
197 a) 16
15.8 15.6 15.4
15.2 15
14.8
14.6
14.4 Maximumsun day in(hours) a 14.2
14 0 50 100 150 200 250 300 350 Dry weight (g/m2)
♦ AC1 AC2
b) 16
15.8
15.6 15.4
15.2
15
14.8 14.6
14.4 Maximum sun in(hours) day a 14.2
14 0 20 40 60 80 100 120 140 160
Dry w eight (g/m2)
♦ AM1 AM2
Figure 3.37 - Scatter plot of maximum daily sun against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
198 a) 600
550
500
450
400
Totalhours of summer sun 350
300 0 50 100 150 200 250 300 350 Dry weight (g/m2)
♦ AC1 AC2
b) 600
550
500
450
400
Totalhours of summer sun 350
300 0 20 40 60 80 100 120 140 160
Dry weight (g/m2)
♦ AM1 AM2
Figure 3.38 - Scatter plot of total hours of summer sun against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
199 600
550
500
450
400 Totalsummer sunhours
350
300 10 10.5 11 11.5 12 12.5 13 Summer mean temperature (oC)
Figure 3.39 - Scatter plot of total summer sun hours against mean summer temperature
Rainfall had a negative (although not significant), relationship with total sun hours
(r=-0.785) as when it is sunny, it is not raining. Figure 3.42 suggests that high rainfall may be associated with low yield. No relationship was found with the summer rainfall data alone however, so it is unlikely that the rain is damaging the flowers directly. High rainfall all at once can lead to waterlogging which is linked with crown rot. If a relationship between maximum rainfall and Arnica exists due to waterlogging, it could be expected that the effect would be more marked for A. montana than A. chamissonis. The former is adapted to well drained, montane environments (Pegtel, 1994) and just occasionally on marshy meadow (Tosco,
1978), whilst the latter is adapted to a broader range of environments (Maguire,
1943) and to wet places (Cronquist, 1955). Figure 3.43 and Table 3.11 seem to suggest this might be the case; although none of the results are significant, the closest is that for A. montana .
200 a) 13.5
13
12.5
12
11.5
11
10.5
Average summer temperature Average (oC)
10 0 50 100 150 200 250 300 350
Dry w eight (g/m2)
♦ AC1 AC2
b) 13.5
13
12.5
12
11.5
11
10.5 Average summer temperatureAverage (oC)
10 0 20 40 60 80 100 120 140 160
Dry weight (g/m2)
♦ AM1 AM2
Figure 3.40 - Scatter plot of mean summer temperature against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
201 a) 13
12.5
12
11.5
11
10.5 Meansummer temperature (oC)
10 0 1000 2000 3000 4000 5000
Flower number
♦ AC1 AC2
b) 13.5
13
12.5
12
11.5
11
10.5
summer Mean temperature (oC)
10 0 500 1000 1500 2000 2500 3000 3500
Flower number
♦ AM1 AM2
Figure 3.41 - Scatter plot of mean summer temperature against flower number for a) AC1 and AC2 and b) AM1 and AM2
202 a) 1300
1200
1100
1000
900
800
Total(mm) for rainfall year 700
600 0 50 100 150 200 250 300 350
Dry w eight (g/m2)
♦ AC1 AC2
b) 1300
1200
1100
1000
900
800 Total (mm) forrainfall year 700
600 0 20 40 60 80 100 120 140 160
Dry w eight (g/m2)
♦ AM1 AM2
Figure 3.42 – Scatter plot of total rainfall against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
203 a) 50
45
40
35
30
25
20 Maximum rainfall (mm) daily 15
10 0 50 100 150 200 250 300 350
Dry weight (g/m2)
♦ AC1 AC2
b) 50
45
40
35
30
25
20 Maximum(mm) dailyrainfall 15
10 0 20 40 60 80 100 120 140 160
Dry w eight (g/m2)
♦ AM1 AM2
Figure 3.43 - Scatter plot of maximum daily rainfall against flower dry weight for a) AC1 and AC2 and b) AM1 and AM2
204
On the whole and as evident in table 3.11, the A. chamissonis plots AC1 and AC2, seemed to have more numerous and more significant relationships with environmental factors, than the A. montana plot AM2. Although this could indicate a lack of sensitivity in trial design, it could also mean that another factor, such as disease, is having more of an impact on yield for this species than the environmental factors and/or that alterations to the quality of the flowers are occurring instead.
3.3.6 Quantitative and qualitative analysis of floral extract
In the 2004, 2005, 2006 and 2007 field seasons, flower material from each of the yield plots was pooled by season, then extracted as described in section 2.2.2.2 and analysed as per sections 2.2.3, 2.2.4 and 2.2.5. The results are presented in
Table 3.12 and Figures 3.44 and 3.46 for A. montana and in Table 3.13 and Figure
3.38 for A. chamissonis . Sesquiterpene lactones below the LOQ were not analysed.
Figures 3.44 and 3.47 both show a clear decline in sesquiterpene lactone production in A.montana from 2004 to 2006 both qualitatively (helenalin and dihydrohelenalin not present in 2006 and 2007) and quantitatively, although total content did increase slightly in 2007. While the total sesquiterpene lactone content of A. montana appears to decline with time (Fig 3.44), that of A. chamissonis remains fairly constant (Fig 3.45 and Table 3.13), although qualitatively there is less diversity, with no helenalin or dihydrohelenalin detectable in 2006 or 2007.
205 Overall, as well as a greater diversity, there are significantly higher levels of sesquiterpene lactones in A. montana extract compared to A. chamissonis extract.
Correlations between the environmental factors and individual sesquiterpene lactone production were investigated and the significant results are presented in table 3.14 for A. montana and table 3.15 for A. chamissonis . (Scatterplots for the
A. chamissonis and A. montana plots were created separately due to the marked difference in content.) It should be kept in mind that the number of repetitions in this case were even less than that for the yield analyses (n=4, df=2 – r crit = 0.950, p<0.05), and yet there were still a number of significant results.
The number of hours of summer sun seemed to have a negative relationship on the levels of individual sesquiterpene lactones in A. montana , in particular methacryl dihydrohelenalin (r (2) =-.967, p<.05), isovaleroyl helenalin (r(2) =-.983, p<.05 (Fig 3.47) and 2-methylbutyryl dihydrohelenalin (r (2) =-.980, p<.05). There was a near significant effect on methacryl helenalin (r=-.933), tigloyl helenalin (r=-
.914), total helenalins (r=-.939) and total sesquiterpene lactones (r=-.927).
Maximum summer wind recorded also had a negative effect on the sesquiterpene lactone content of AM2, in particular on acetyl helenalin (r (2) =.963, p<.05).
206
9.00
8.00 7.00 6.00
5.00
4.00
3.00
Sesquiterpene lactonesSesquiterpene (mg/g) 2.00
1.00 0.00 AM1 AM2 AM2 AM2 AM2 2004 2005 2006 2007
Total Helenalins (mg/ml) Total Dihydrohelenalins (mg/ml) Total Sesquiterpene lactones (mg/ml)
Figure 3.44 –Total sesquiterpene lactone content for AM1 and AM2 over a four year period
AM1 material was not available for years other than 2004
3.00
2.50
2.00
1.50
1.00
Sesquiterpene(mg/g) lactone 0.50
0.00 AC1 AC2 AC1 AC2 AC1 AC2 AC1 AC2 2004 2005 2006 2007 Dihydrohelenalin Helenalin 4-O-Acetyl-6-desoxychamissonolide Chamissonolide Arnifolin Total sesquiterpene lactone content
Figure 3.45 –Total sesquiterpene lactone content for AC1 and AC2 over a four year period
207
Sesquiterpene lactones (mg/g) 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
Isovaleroyl dihydrohelenalinIsovaleroyl Isobutyryl dihydrohelenalin Dihydrohelenalin Figure 3.46 – Individual sesquiterpene cont 3.46Figure – lactone Individual *
M M M M AM2 AM2 AM2 AM2 AM1 AM1 material was not available for years other than other years for available not was material AM1 *
* 2-methylbutyryl dihydrohelenalin Methacryl helenalin Helenalin 2004 *
*
*
*
Tigloyl helenalinTigloyl Methacryl dihydrohelenalin Acetyl helenalin * * * ent of extracts from AM1 and AM2 over afour overyear p and AM2 entfrom of extracts AM1 2004. The symbol * indicates values below the LOQ LOQ the below values indicates * The symbol 2004.
0520 2007 2006 2005 *
* * Tigloyl dihydrohelenalinTigloyl helenalinIsovaleroyl Acetyl dihydrohelenalin *
*
*
2-methylbutyryl helenalin Isobutyryl helenalin *
* *
* eri od
* *
208
Table 3.12 – Sesquiterpene lactone content (mg/ml) of AM1 and AM2 over a four year period
2004 2005 2006 2007 Sesquiterpene lactone AM1 AM2 AM1 AM2 AM1 AM2 AM1 AM2 (mg/ml) Dihydrohelenalin 0.03 0.04 0.03 0.00 0.00
Helenalin 0.11 0.12 0.05 0.00 0.00
Acetyl helenalin 1.43 1.58 1.36 0.64 0.80
Acetyl dihydrohelenalin 0.13 0.09 0.14 0.19 0.06
Isobutyryl helenalin 1.72 1.72 1.64 0.89 1.30
Isobutyryl dihydrohelenalin 0.06 0.04 0.06 0.04 0.05
Methacryl helenalin 1.63 1.53 1.40 0.83 1.17
Methacryl dihydrohelenalin 0.14 0.10 0.11 0.02 0.07
Isovaleroyl helenalin 1.18 1.08 1.33 0.34 0.92
2-methylbutyryl helenalin 0.53 0.43 0.45 0.12 0.31
Isovaleroyl dihydrohelenalin 0.03 0.02 0.03 0.00 0.01 2-methylbutyryl 0.11 0.10 0.06 0.00 0.01 dihydrohelenalin Tigloyl helenalin 0.90 0.74 0.77 0.27 0.45
Tigloyl dihydrohelenalin 0.05 0.03 0.05 0.02 0.00
Total Helenalins 7.48 7.21 7.01 3.07 4.95
Total Dihydrohelenalins 0.55 0.43 0.48 0.29 0.20 Total Sesquiterpene 8.03 7.64 7.48 3.36 5.14 lactones Total Sesquiterpene 2 569.9 697.5 202.1 52.8 54.0 lactones (mg/m )
AM1 material was not available for years other than 2004. Figures in italics are below LOQ (0.08mg/ml)
209 Table 3.13 – Sesquiterpene lactone content (mg/ml) of AC1 and AC2 over a four year period
2004 2005 2006 2007 Sesquiterpene lactone AC1 AC2 AC1 AC2 AC1 AC2 AC1 AC2 (mg/ml)
Dihydrohelenalin 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00
Helenalin 0.03 0.04 0.03 0.04 0.00 0.04 0.00 0.00 4-O-Acetyl-6- 0.17 0.15 0.11 0.09 0.13 0.15 0.18 0.18 desoxychamissonolide Chamissonolide 1.61 1.63 1.41 1.58 1.82 1.89 1.87 1.85
Arnifolin 0.31 0.32 0.24 0.15 0.29 0.32 0.34 0.34 Total sesquiterpene lactones 2.13 2.15 1.80 1.88 2.25 2.39 2.39 2.36 (mg/g) Total sesquiterpene lactones 157.8 146.1 78.3 113.3 196.1 177.3 77.1 72.8 (mg/m 2)
Table 3.14 - Correlations of the AM2’s sesquiterpene lactone content with climate data
Environmental factor Summer Summer No. of gale Sesquiterpene lactone Max wind sun hours max wind days Acetyl helenalin -0.796 -0.963 0.438 0.877
Acetyl dihydrohelenalin 0.684 0.156 -0.056 -0.030
Isobutyryl helenalin -0.947 -0.833 0.550 0.835
Methacryl helenalin -0.933 -0.845 0.481 0.796
Isovaleroyl helenalin -0.983 -0.599 0.781 0.824
2-methylbutyryl helenalin -0.980 -0.750 0.646 0.837
Tigloyl helenalin -0.914 -0.843 0.661 0.928
Total helenalins -0.939 -0.841 0.592 0.872
Total dihydrohelenalins -0.516 -0.832 0.600 0.948
Total sesquiterpene lactones -0.927 -0.854 0.601 0.891
Correlation (r crit =0.950, p<0.05) Not significant but above 0.8 Not significant
210 Table 3.15 - Correlations between AC1 and AC2’s sesquiterpene lactone content with climate data
Max. Sun No. Of gale Environmental factor Max. Wind Max.Temp hours days Sesquiterpene AC1 AC2 AC1 AC2 AC1 AC2 AC1 AC2 lactone 4-O-Acetyl-6-desoxy- -0.567 -0.819 -0.966 -0.744 -0.451 -0.795 0.701 0.844 chamissonolide
Chamissonolide -0.816 -0.697 -0.413 -0.105 -0.988 -0.984 0.472 0.222
Arnifolin -0.737 -0.944 -0.848 -0.637 -0.733 -0.793 0.762 0.734
Total Sesquiterpene -0.838 -0.890 -0.598 -0.347 -0.929 -0.976 0.604 0.550 lactones
Correlation (r crit =0.950, p<0.05) Not significant but above 0.8 Not significant
1.60
1.40
1.20
1.00
0.80
0.60 Isovaleroyl helenalin Isovaleroyl 0.40
0.20
0.00 350 400 450 500 550 600
Total summer sun hours
Figure 3.47 - Correlation of isovaleroyl helenalin against total summer sun hours in AM2
211
For AC1 and AC2, the number of gale days seems to have had a significant negative effect on levels of chamissonolide in both AC1 (r (2) =-.988, p<.05) and
AC2 (r (2) =-.984, p<.05) (Figure 3.48), and on the total content of sesquiterpene lactones ((r (2) =-.929 and -.976 respectively, p<.05).
Figure 3.48 - Correlation of number of gale days against chamissonolide AC1 and AC2
14.5
14
13.5
13
12.5
12
11.5 Numberof days gale 11
10.5
10 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Chamissonolide (mg/ml)
♦ AC1 AC2
The relationship between sesquiterpene lactone content and climate seems to be complex, so to investigate any patterns in sesquiterpene lactone production further, the extracts from AM2, AC1 and AC2 flower material were tracked over two field seasons (2006 and 2007). The results for A. montana are presented in tables 3.16 and 3.17, respectively, and those for A. chamissonis in tables 3.18 and
3.19. In the case of A. montana the highest level of sesquiterpene lactones
212 (mg/ml) were produced from material harvested at the start of the season. The early flowers were normally the main flowers for each plant. These were the heaviest and it may be the plants invest more in defending the largest and showiest of the flowers. Pollination of these would allow early seed set which could be an advantage in an environment where early frosts are common.
Table 3.16 Sesquiterpene lactone (SL) content of AM2 produced in 2006
AM2 - 2006
20/ 06 30/ 06 07/ 07 14/ 07 21/ 07 29/ 07 04/ 08
Total Helenalins (mg/ml) 4.26 2.95 3.67 4.04 2.19 2.30 0.93 Total Dihydro-helenalins 0.12 0.07 0.07 0.54 0.10 0.43 0.32 (mg/ml) Total SLs (mg/ml) 4.37 3.02 3.74 4.58 2.29 2.73 1.25
Dry flower weight (g/m 2) 1.37 4.19 14.66 24.41 18.70 2.24 0.27
Dry flower weight (g/m 2) 0.33 1.00 3.49 5.81 4.45 0.53 0.06
Total number of flowers 5.00 15.00 64.00 157.00 124.00 16.00 2.00
Dry weight per flower 0.27 0.28 0.23 0.16 0.15 0.14 0.14
Total SL content (mg/g) 4.37 3.02 3.74 4.58 2.29 2.73 1.25 Total SLs produced (g/ 1.43 3.01 13.06 26.60 10.20 1.46 0.08 m2)
213 Table 3.17 Sesquiterpene lactone (SL) content of AM2 produced in 2007
AM2 – 2007
15/ 06 21/ 06 28/ 06 04/ 07 09/ 07
Total Helenalins (mg/ml) 7.64 3.82 4.73 3.61 4.93
Total Dihydro-helenalins (mg/ml) 0.27 0.14 0.16 0.15 0.20
Total SLs (mg/ml) 7.97 3.97 4.89 3.77 5.13
Dry weight (g/m 2) 11.83 7.95 13.45 8.84 1.86
Dry weight (g/m 2) 2.82 1.89 3.20 2.10 0.44
Total number of flowers 42.00 36.00 85.00 65.00 15.00
Dry weight per flower 0.28 0.22 0.16 0.14 0.12
Total SL content (mg/g) 7.97 3.97 4.89 3.77 5.13
Total SLs produced (g/ m 2) 22.44 7.51 15.65 7.93 2.27
However, the pattern for A. montana seems to vary a great deal both from year to year and within the season as illustrated in figures 3.49 and 3.50. In 2006 the sesquiterpene lactone production of A. montana decreased as the season progressed, whilst in 2007 it started very high (8.0mg/ml) and then dropped to between 4-5.0 mg/ml for the rest of the season. This could be due to decreased sun hours in July in 2007 (a drop from 156.5 hours in June to to 91.2 hours in July) whilst in 2006, a later start to the growing season resulted in the plants flowering into the start of August which may have been caused by the a sunnier June and
July (157 and 173 hours of sun respectively). The peak point of production
(sesquiterpene content multiplied by dry weight), was the 14 th of July for 2006 and
15 th of June in 2007. This suggests that finding an optimum time for harvesting could be difficult.
214
9.00 120.00
8.00 100.00 7.00
6.00 80.00 5.00 60.00 4.00
3.00 40.00
2.00 20.00 Sesquiterpene lactone Sesquiterpene (mg/ml) 1.00 0.00 0.00 20/06/2006 30/06/2006 07/07/2006 14/07/2006 21/07/2006 29/07/2006 04/08/2006 Total sesquiterpenelactones inharvest (mg) AM2
Total helenalins (mg/g) Total dihydrohelenalins (mg/ml) Total sesquiterpene lactones (mg/g) Total sesquiterpene lactones in harvest (mg)
Figure 3.49 - Sesquiterpene lactone production in A. montana over the 2006 field season
9.00 120.00 8.00 100.00 7.00
6.00 80.00
5.00 60.00 4.00 3.00 40.00
2.00 20.00 Sesquiterpene lactone (mg/ml) lactone Sesquiterpene 1.00 0.00 0.00 15/06/2007 21/06/2007 28/06/2007 04/07/2007 09/07/2007 AM2 sesquiterpeneTotal lactones in harvest(mg)
Total Helenalins(mg/g) Total Dihydrohelenalins (mg/g) Total Sesquiterpene Lactones (mg/g) Total sesquiterpene lactones in harvest (mg)
Figure 3.50 - Sesquiterpene lactone production in A. montana over the 2007 field season
215 Table 3.18 Sesquiterpene lactone (SL) content of AC1 and AC2 produced in 2006
AC1 - 2006 AC2 - 2006 07/ 14/ 21/ 28/ 04/ 07/ 14/ 21/ 28/ 04/
07 07 07 07 08 07 07 07 07 08 Dihydro- 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 helenalin Helenalin 0.00 0.00 0.00 0.00 0.02 0.01 0.00 0.01 0.02 0.17 4-O-Acetyl-6- desoxy- 0.12 0.11 0.10 0.12 0.19 0.11 0.14 0.14 0.14 0.19 chamissonolide Chamissonolide 1.78 1.71 1.46 1.66 2.48 1.78 1.82 1.73 1.78 2.33
Arnifolin 0.25 0.24 0.20 0.30 0.47 0.28 0.27 0.26 0.32 0.44 Dry weight 7.51 48.93 28.31 25.06 21.19 8.47 39.71 18.91 21.49 22.5 (g/m 2) Dry weight 2 5.01 32.62 18.87 16.71 14.13 5.65 26.47 12.61 14.33 15.00 (g/m ) Total number of 92 702 456 395 407 90 563 302 329 310 flowers Dry weight per 0.08 0.07 0.06 0.06 0.05 0.09 0.07 0.06 0.07 0.07 flower Total SL content 2.15 2.06 1.77 2.08 3.17 2.19 2.24 2.15 2.26 3.13 (mg/g) Total SLs 10.8 67.3 33.4 34.8 44.7 12.3 59.3 27.0 32.4 47.0 produced (g/m2)
The opposite relationship appears to be the case for A. chamissonis (Figs 3.51-
3.54). In both AC1 and AC2 in 2006, sesquiterpene lactone production increased towards the end of the season and in the less sunny month of August. However, in
2007 the yield of sesquiterpene lactones was fairly constant, perhaps declining slightly towards the end of the season in AC1. The point of maximum production was the 14 th of July for AC1 and AC2 in 2006 and between the 6 th and the 14 th of
July in 2007. This suggests that mid season would be a reasonable time to harvest
A. chamissonis, but at least another season would be required to say with any certainty that this is the case.
216 Table 3.19 Sesquiterpene lactone content of AC1 and AC2 produced in 2006
AC1 - 2007 AC2 – 2007
28/06 06/07 15/07 23/07 28/06 06/07 15/07 23/07
Dihydrohelenalin 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Helenalin 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 4-O-Acetyl-6-desoxy- 0.18 0.19 0.14 0.20 0.14 0.19 0.15 0.23 chamissonolide Chamissonolide 1.98 2.10 1.64 1.75 1.74 2.00 1.66 2.00
Arnifolin 0.42 0.42 0.26 0.26 0.35 0.42 0.28 0.30
Dry weight (g/m 2) 3.13 15.27 21.13 8.85 3.62 16.22 19.83 6.56
Dry weight (g/m 2) 2.09 10.18 14.09 5.90 2.41 10.81 13.22 4.37
Total number of flowers 40 202 343 192 43 226 323 141
Dry weight per flower 0.08 0.08 0.06 0.05 0.08 0.07 0.06 0.05
Total SL content (mg/g) 2.58 2.72 2.04 2.21 2.22 2.61 2.09 2.53
Total SLs produced (g/ m 2) 5.4 27.7 28.7 13.0 5.4 28.2 27.6 11.1
3.50 120.00
3.00 100.00
2.50 80.00 2.00 60.00 (mg)
(mg/ml) 1.50
40.00 1.00
0.50 20.00 Total sesquiterpeneTotal lactones inharvest (mg) Amount of sesquiterpene lactone lactone sesquiterpene of Amount 0.00 0.00 Total lactonessesquiterpene harvest per 07/07/2006 14/07/2006 21/07/2006 28/07/2006 04/08/2006 AC1
Dihydrohelenalin Helenalin 4-O-Acetyl-6-desoxychamissonolide Chamissonolide Arnifolin Total sesquiterpene lactones (mg/ml) Total sesquiterpene lactones per harvest (mg)
Figure 3.51 - Sesquiterpene lactone production in AC1 over the 2006 field season
217 3.50 120.00
3.00 100.00 2.50 80.00 2.00 60.00 (mg)
(mg/ml) 1.50 40.00 1.00 0.50 20.00
lactone sesquiterpene of Amount 0.00 0.00 Total lactones sesquiterpene harvestper Total sesquiterpeneTotal lactones inharvest (mg) 07/07/2006 14/07/2006 21/07/2006 28/07/2006 04/08/2006 AC2
Dihydrohelenalin Helenalin 4-O-Acetyl-6-desoxychamissonolide Chamissonolide Arnifolin Total sesquiterpene lactones (mg/ml) Total sesquiterpene lactones per harvest (mg)
Figure 3.52 - Sesquiterpene lactone production in AC2 over the 2006 field season
3.50 120.00
3.00 100.00
2.50 80.00 2.00 60.00 (mg)
(mg/ml) 1.50 40.00 1.00
0.50 20.00 Amount of sesquiterpene lactone lactone sesquiterpene of Amount Total lactonessesquiterpene harvestper
0.00 0.00 sesquiterpeneTotal lactones in harvest(mg) 28/06/2007 06/07/2007 15/07/2007 23/07/2007
AC1
Dihydrohelenalin Helenalin 4-O-Acetyl-6-desoxychamissonolide Chamissonolide Arnifolin Total sesquiterpene lactones (mg/ml) Total sesquiterpene lactones per harvest (mg)
Figure 3.53 - Sesquiterpene lactone production in AC1 over the 2007 field season
218
3.50 120.00
3.00 100.00 2.50 80.00 2.00 60.00 (mg)
(mg/ml) 1.50 40.00 1.00
0.50 20.00 Amount of sesquiterpene lactone lactone sesquiterpene of Amount Total sesquiterpene lactones harvestper
0.00 0.00 sesquiterpeneTotal lactones inharvest (mg) 28/06/2007 06/07/2007 15/07/2007 23/07/2007 AC1 Dihydrohelenalin Helenalin 4-O-Acetyl-6-desoxychamissonolide Chamissonolide Arnifolin Total sesquiterpene lactones (mg/ml) Total sesquiterpene lactones per harvest (mg)
Figure 3.54 - Sesquiterpene lactone production in AC2 over the 2007 field season
3.4 Discussion of the yield and floral extract results
The aim of this chapter was to determine the yield of A. montana and A. chamissonis over several years and to determine the extent to which environmental factors influence yield in both qualitative and quantitative terms.
Whilst it is important that the yield of flowers is enough to make the crop financially sustainable, it is also important that the sesquiterpene lactones within the flowers are of the quantity and type required for medicinal purposes. This information can then be used to identify characteristics that future sites should possess, identify the best time for harvesting, and potentially aid in the development of agronomic regimes that maximise both the yield and the quality of Arnica produced.
219
The literature has reported yields of Arnica ranging from 94 to 284 kg per hectare in Italy (Bezzi & Ghidini, 1989), 640 kg per hectare in Switzerland (Delabays and
Mange, 1991) to over 1000 kg/ha in both Germany (Bomme & Daniel, 1994) and
Finland (Galambosi et al. 1998). In the first year of this study, extrapolated yield corresponded to 1407 kg/ha for A. montana and 1418 kg/ha for A. chamissonis , more than other reported results and which was evidence of Arnica’s potential to be a high yielding crop in the Orkney environment. However, this yield dropped dramatically to 105 kg/ha for A. montana and 323-308 kg/ha for A. chamissonis by
2007. Data are not available regarding the long term yield of the trials in the rest of
Europe, but a general decline in yield has been reported (Cassells et al, 1999).
Hence either a method of agronomy that allows Arnica to be more sustainable over the longer term is required, or the factors implicated in this decline should be identified to allow other potential sites to be characterised.
As outlined earlier in this chapter, soil is crucial for effective land use. In agriculture, it has a strong influence on yield through its control over the rhizosphere of the plant. Whilst soil characteristics tend to be interrelated, and hence correlations of individual characteristics and yield are difficult, soil texture is one of the more stable characteristics and hence the most influential (Ellis &
Mellor, 1995). It was determined that the soil in this area was relatively high in clay
(40-50%). High clay content can decrease pore size and lead to mechanical impedance to the plant root as well as problems with water regime, nutrient flow and aeration (Killham, 1994). The silty clay soil also contained hydrated oxides of iron which suggested that waterlogging was common. As A. montana is more
220 often found on well drained podzols (Pegtel, 1994), it is possible this could be the cause of a decline in yield for this species. A. chamissonis on the other hand, is more often found in wet places (Rickett, 1973), but is also found on alkaline soil and this site was determined to be pH 6.0. In a study of niche modelling, Parolo et al (2008) found that in the alpine environment, habitat, elevation and geomorphology had the highest predictive power of natural populations. This would seem to give weight to the argument that soil is a strong influencer of the yield of Arnica in Orkney.
Whist the soil aggregation was good, the high clay content is likely to have impeded drainage and maintained high levels of the micro-organisms
Phytophthora and Pythium, the pathogens suspected of causing the crown rot in
A. montana (DEFRA, 2005). Both of these pathogens have been described as being favoured by waterlogging (ibid), which could in turn, have led to the high levels of disease evident in the AM1 plot. Whilst agronomic regimes such as weeding may have potentially damaged the plants and left them prone to infection
(see Chapter 4), it could also be that environmental factors such as heavy rain increased the plant’s vulnerability to disease. Although the data didn’t show any significant negative relationships between yield and rainfall, the results obtained were stronger for A. montana than they were for A. chamissonis .
Only the flowering stems of A. montana are exposed to the winds to any great extent and hence it could be expected to be less susceptible to strong winds than
A. chamissonis . This is reflected in the data, with only the yield of A. chamissonis negatively correlated with average summer wind speed. The negative correlation
221 applied to both the sheltered and unsheltered plants and it may be that the screen provides very little benefit for this plant. Although a relationship between low yield and wind would at first suggest damage has occurred to the plant (stem breakages etc), the number of gale days and maximum wind speed had very little obvious effect.
Wind did however, seem to have a negative impact on the levels of sesquiterpene lactones both in A. chamissonis and A. montana, which could be a sign of resources diverted for repair. Although lignins are produced by the shikimic acid pathway and terpenes from the mevalonate pathway, both have phosphoenol pyruvic acid as a common precursor (Mooney, 1972). It is perhaps the case that in more consistently windy conditions, A. chamissonis is stimulated to invest more in plant structures such as in the stems or the roots, instead of the production of sesquiterpene lactones in the flowers. Further investigations of the total plant mass and composition of the plots would be required to investigate this further, which was not possible, due to insufficient plants for destructive harvesting.
Although there was a correlation with average wind speed and increased rainfall, none of the plots showed any clear sign of a relationship between rainfall and yield.
As the only benefit of the wind break seemed to be a slightly increased yield for A. chamissonis, it was decided no new plots should be screened. The next stage of trials were all conducted on fresh ground to avoid existing disease areas, and either up slope from the yield plots, or on raised furrows.
222 There did seem to be some evidence of a positive relationship between maximum number of sun hours and the yield of A. chamissonis. This seemed to be strongest out with the months of May – June and it may be that the period after flowering is important for the root reserve to be replenished for the following year as it is for grapes (Quirk & Somers, 2009). Again however, further work involving total harvests of whole plants would be required to confirm whether this was the case.
In contrast, a negative relationship was in evidence for the maximum hours of summer sun and the levels of sesquiterpene lactones in A. chamissonis, as well as for the total summer hours of sun and extract content for A. montana. If physical yield was correlated with increased sun, it could indicate that the sesquiterpene lactones were being diluted in the increased physical mass of flower material. However, summer sun had no significant effect on either flower dry weight or flower number for the AM2 plot. Instead, it may be that increased sun has lead to an increase in plant biomass (leaves etc), and the diversion of resources to other plant parts during the period of peak vegetative growth.
Schmidt et al. (1998) found that young shoots of A. montana accumulated mainly helenalin derivatives, which then decreased to near zero at the onset of leaf formation (3-4 weeks). The authors suggest this is best explained by the presence of a hydrogenase system that converts the sesquiterpene lactones from the hydro to the dihydro type. Further investigation into this pathway would be of interest as it may be that leaf formation and growth results in the diversion of resources away from defence. This would require additional whole plant analysis however, and is beyond the scope of this study.
223
Scheidel & Bruelheide (1999) found that the palatability of A. montana leaves decreased after damage had already occurred. This suggests either the mobilisation of defence compounds that were not otherwise present in this area of the plant, or the rapid increase in production of these compounds in situ . If these defence compounds are the sesquiterpene lactones under investigation in this study, then it may be that the decrease in sesquiterpene lactones associated with increased hours of sun reflects a decreased need for defence as the drier conditions make the surrounding environment less suitable for slugs and snails.
However, it could then be expected that yield of the compounds would increase with increased rainfall, and no indications of such a relationship were found.
Although vernalisation was thought more likely to be a requirement for A. montana, both species showed evidence of a relationship between increased wind chill and lower temperatures and increased yield. If winters in Orkney become milder as climate models seem to predict (e.g. Kerr et al. 1999), then it may be that future sites for A. montana trials should be based in a more montane environment. Again, colder winters and wind chill are likely to reduce the herbivore population and so this may be the more influential factor. Indeed, Bruelheide and
Scheidel (1999), found very little herbivore damage when A. montana was grown in its natural habitat (610 metres above sea level); they attributed this to the lack of slugs and snails in the surrounding environment.
The trials in Orkney revealed some important data on ideal harvesting times. A. chamissonis flowers over a shorter period whilst the flowering time for A. montana
224 is more staggered. This may be because A. montana is found more on high alpine slopes (Maguire, 1943 and Pegtel 1994), where long staggered flowering would be an advantage when late frosts occur (Forbes & Watson, 1992). This would not be as important for A. chamissonis as it is found more often in valleys (Cronquist,
1955).
Whilst the data for both A. montana and A. chamissonis revealed that levels of sesquiterpene lactones declined from the start of the season to the end, the total harvest levels for A. montana were variable throughout. Consequently, if both dry weight and sesquiterpene lactone content are to be maximised, then constant hand harvesting would be required. A. chamissonis on the other hand had a peak harvest level that was the same in 2006 as it was in 2007 which may indicate that mechanised harvesting would be more effective for this species.
Although production was high in the first year, the yield data suggests that A. montana is not as suited to the Orkney environment as A. chamissonis . Whilst A. montana produces a greater amount and number of sesquiterpene lactones, their initially very high levels may be indicative of the plants being under extreme stress and / or dying. Total sesquiterpene lactone content for A. montana decreased in each year of this study, whilst those of A. chamissonis remained fairly constant.
Certainly the A.chamissonis plants seem more influenced by environment than A. montana and this may indicate that for the latter, resources are being diverted elsewhere (e.g. defence).
225 Parolo et al (2008) found that soil wetness, solar radiation, slope angle and slope aspect had very little correlation with natural populations and it is of course possible that all of the correlations in this chapter were artefacts of the low numbers of repetitions, or the by-product of one environmental factor being interrelated with another. Regardless, the yield plots have revealed that alternative agronomic regimes are required for A. montana if yields are to be increased.
226
4. The influence of fertiliser and weeding on yield
4.1 Introduction
“In both natural and agricultural communities, the environment is seldom optimal
for plant growth.”
F Stuart Chapin III, 1991, p21
All mesic environments are subject to variations in light, temperature, nutrients and moisture. Most natural environments not currently under agriculture are deficient in respect to at least one of these parameters, yet remain of interest, as they present potential areas of agricultural expansion (Chapin, 1991). To identify crops suitable for these environments, a clear understanding is needed of whether the selected plant will adapt to the unique conditions of the site and any consequences such adaptations might have on yield. The previous chapter has provided indications as to how both yield and quality vary due to environmental conditions, but further investigation is required to determine how Arnica responds to variations in agronomic regime.
As outlined in the previous chapter, the harvest index of a crop can vary considerably in both spatial and temporal terms. The weather is one of the most important influencing factors in this variation, but another is biological e.g. weeds, pests and pathogens. Normally the latter are unwanted, as they have the potential to reduce the harvest index and contaminate the product. Crops that require frequent weeding are both more expensive to grow (particularly if hand weeding is required) and susceptible to root damage as the encroaching weeds are removed.
227 Weeds can also serve as a reservoir of crop pests, particularly where the weed’s growing season is longer than that of the crop (Tivy, 1990), and can indirectly decrease the leaf area or effectiveness by shading. Weeds are a recurrent and nearly universal threat to crop productivity (Liebman & Gallandt, 1997) and so their management is important.
Since the early 1960s, the focus of farmers in terms of weed control has been almost entirely on herbicides, with the total world herbicide market worth the equivalent of about £9,180 million in 1995 (British Agrochemicals Association,
1996). This increase has corresponded to increased yields and productivity
(Liebman & Gallandt, 1997), but also an increased awareness of: ground and water contamination (Leistra & Boesten, 1989), the health risks associated with herbicide application (WHO, 1990) and increasing numbers of herbicide resistant weeds (Vyvyan, 2002). These factors have in turn led to the increasing popularity of alternative approaches to weed control, such as “ecological weed management”
(Liebman & Gallandt, 1997).
Ecological weed management is of particular interest when cultivating crops such as Arnica as the market for this crop is primarily for flowers grown organically
(Burnie, unpublished), a regime in which the use of herbicides is not permitted.
The aim of ecological weed management is to place stresses on weeds to reduce the:
a) density of their propagules and seedlings
b) rate of weed seedling emergence compared to crop seedling
emergence
228 c) proportion of resources consumed by weeds
d) rate of weed dispersal within and between fields and
e) proportion of particularly problematic genotypes and species
(Aldrich, 1984).
Techniques that can be employed include; increased diversity of crop vegetation
(intercropping, green manuring etc), soil disturbance through tillage and cultivation, herbivory, disease (either applied or indigenous), variations in crop density and arrangement, manipulations of environmental conditions (eg. liming), and the application of herbicides (the latter only to be used when other systems fail (Liebman & Gallandt, 1997)). Unfortunately, such methods are difficult to implement on the small trial scale.
Allelopathy is another important potential tool for organic weed control, as it involves the production and release of allelochemicals from the flowers, leaves, seeds and roots of plant materials (Weston, 1996). For example, Eucalyptus camaldulensis Dehnh. essential oil, with spathulenol as the main active compound, completely inhibited the germination and growth of the weeds
Amaranthus hybrdius L. and Portulaca oleracacea L. (Verdeguer et al. 2009).
Spathulenol was shown in this research (section 2.3.2) to be a strong component of the A. chamissonis floral essential oil (section 2.3.3).
Artemisia nova A.Nels (Asteraceae) is another example of a plant that is believed to produce germination and growth inhibiting compounds and like Arnica it has glandular trichomes containing sesquiterpene lactones. Kelsy & Shafizadeh.
229 (1980) found that soil surrounding sagebrush contained sesquiterpene lactones and terpenes in similar proportions to that of the leaves. Germination of sage seed was inhibited when grown on sage litter, which when analysed, was also found to contain both terpenes and sesquiterpene lactones. The majority of the inhibition was believed by the authors to be due to the terpenes, though Macias et al. (2000) found that the guaianolide sesquiterpene lactone (see section 1.9.3.1) dehydrozaluzanin was a potent plant growth regulator when levels were above
100 µM. As sesquiterpene lactones are not volatile, it is believed that when they are present in the soil it is due to leaf drop and direct leaching. If similar levels of sesquiterpene lactones accumulate in the soil around the Arnica plants, then this could provide a natural form of weed control.
A. chamissonis was found to not require weeding and it could be the higher competitiveness of this species is due to a natural allelopathy rather than its growth habit (section 3.3.4). Either way, investigations of the effectiveness of weeding treatment were not required for this species.
The sesquiterpene lactone content of the leaves of A. montana is not normally investigated as it is only the flowers that are of commercial interest. Schmidt et al.
(1998) studied the sesquiterpene lactone content of field and in vitro A. montana and found that although the early leaves of field grown plants have a high content of helenalin type sesquiterpene lactones, these convert to mainly dihydro types after 3-4 weeks, which was also found by Poplawski et al. (1971). It is possible that sesquiterpene lactones are only produced at these levels when the plants are young as that is when they are most susceptible to herbivory. Regardless, the
230 dihydro types have been shown to be less effective anti-inflammatory agents
(Wagner et al. 2004b) which may explain the lack of commercial interest in leaf extract.
Although it was not possible to isolate individual sesquiterpene lactones of either form to the levels required to determine their effectiveness as plant growth inhibitors, if the leaching of compounds from A. montana is effective at reducing the encroachment of the surrounding vegetation - regardless of the compound responsible – the effect should be detectable by simply not weeding and monitoring the plots.
The production and storage of secondary compounds such as allelochemicals are energy intensive processes (Levin, 1976). As stressed by Reichardt et al. (1991), metabolite levels in response to the environment, are very much dependent on the dynamics associated with their production and turnover and the resources available to the plant. Chapter 3 investigated some of the effects of the environment (e.g. sun, rain, wind) but mineral supply is another important factor for consideration.
Nitrogen, phosphorous and potassium are the three elements commonly required to be added to soil in agricultural regimes and are normally available to the plant in
+ - 3- + + the form of ions: NH 4 or NO 3 , PO 4 and K . Of these, it is the capture of NH 4
- and NO 3 which is among the most energy intensive of the activities the plant can
+ - undertake (Bloom, 1997). NH 4 displays less seasonal variation than NO 3 and so it tends to be the preferred source of nitrogen for plants, although most rely on
231 both for growth. Nitrate is highly soluble and can be accumulated without toxicity
+ - effects (Lorenz, 1978). Seedlings however tend to absorb more NH 4 than NO 3
+ when present in equal amounts, possibly due to the uptake of NH 4 requiring less energy. However, in well aerated temperate agricultural soils the availability of
- + NO 3 can outweigh NH 4 by an order of magnitude (Bloom, 1997).
In a study of plant growth rate and herbivore defence, Coley (1988) found that the more resources a plant allocated to growth, the less it invested in defences such as tannins and lignins. So when the availability of resources such as nitrogen are limiting (and growth is restricted), there can be an associated increase in synthesis of secondary metabolites via the accumulation of resources that are not limited, for example, carbon (Bryant et al. 1983 & 1995). This the basis of the carbon/nutrient balance hypothesis which predicts that concentrations of secondary metabolites change in response to changes in light and mineral nutrition. Hence in sunny conditions and where nutrients are limited, when carbohydrate accumulates in excess of growth demands, the excess carbon could be diverted to produce secondary metabolites such as tannins (Chapin, 1991).
Sesquiterpene lactone production has been shown to be induced by herbivore attack in A. montana and can cause enzyme deactivation in animals and insects
(Picman, 1986 and Scheidel & Bruelheide, 1999). If it is presumed that helenalin, dihydrohelenalin and their esters in A. montana and A. chamissonis serve as a form of defence, then it would reduce the ‘cost’ of production if such compounds were either a) only produced and stored when other factors (such as low nutrient
232 availability) had reduced the growth rate of the plant, of b) if production was only triggered after damage had occurred.
Almeida-Cortez et al. (2004) found in a study of six Asteraceae species that phytochemical toxicity increased with decreasing nutrient levels, indicating that the plants with the highest level of secondary metabolites were those whose growth was most limited. Carbon supply does not limit plant development when nitrate is limited, so increased levels of carbon-based defensive compounds would provide a selective advantage (Bryant et al. 1983). On the other hand, where the plant is subjected to low carbon uptake or high respiration (low light or in the case of disease) then less resource may be allocated to the production of defensive compounds. Almeida-Cortez et al. (2004) also found in their study that the strongest effect was found where irradiance was also limited although they found species-specific differences.
Higher levels of monoterpene and sesquiterpene lactones have been demonstrated in plants grown under 'nitrogen-deficient' conditions compared to those grown in nitrogen-rich soils (Mihaliak & Lincoln, 1989). Gartlan et al. (1980) compared levels of defensive compounds in forests on two different sites and found that the forest on the more nutrient rich soil had lower levels of these compounds. Arnica is normally found on low nutrient soils and so it could be predicted that it would invest in higher levels of defensive compounds to avoid the losses associated with herbivory. It may be when this species is grown in a more nutrient rich environment, resources aren’t allocated towards defence until after herbivory has occurred. It may be that the internal economics of Arnica lead to
233 greater investment in structure when nutrients are in excess, rather than in defence, leading to decreased levels of the active compounds. This would be of selective advantage, as if Arnica is growing in a more fertile soil, weeds are likely to be more competitive for resources such as light and water.
A strong selective pressure exists for species that can store resources and then mobilise them when they undergo competition for resources. Where defensive compounds are used solely for defence, then their cost is a direct reduction in resources related to fitness. For example, if the sesquiterpene lactones of Arnica serve only as defensive compounds as hypothesised and if their production occurred regardless of whether nutrients were limiting, there would be a cost to the potential for growth of the plant if the standard cost/benefit model is to be invoked
(Eissenstat, 1997).
Nutrient levels can also have consequences for both quality and quantity of essential oil produced. For example coriander produced less essential oil when nitrogen was plentiful (de la Fuente et al. 2003), and Azizi et al. (2008) found a decrease in essential oil production of three oregano populations with increased nitrogen (which they explained by the dilution effect of an increase in dry matter), whilst percentage content of individual compounds was unaffected.
Where inhibition of weeds by allelopathy is not sufficient but where weeds need to be removed, there is a risk of physical disturbance caused by removing the weeds close to the A. montana plants. This can lead to damage of the stem and / or roots and such damage could in turn cause the plant to be more susceptible to crown rot
234 or similar diseases with a corresponding increase in plant deaths. To study this further, investigations into the effects of weeding treatment, on yield and plant number were conducted for this species.
When designing agronomic treatments for Arnica, it is important a balance is struck between creating a stress in order to stimulate production, and depriving the plant of enough resources to increase production of secondary compounds in order to defend against attack or repair damage. Whilst inducible responses and flexible defence strategies are of benefit to the plant, they can make attempts to simultaneously maximise flower quality and increase flower yield difficult. For example, it would considerably reduce the cost of Arnica production if it did not require weeding or fertiliser, as both chemical additives and labour are expensive.
It may also be the case that a decrease in yield resulting from not applying fertiliser would be outweighed by the increased value of the extract, or that the costs of weeding control treatments would outweigh the benefits in terms of increased flower yield.
Conversely, if it is shown that the reduction in the yield of flowers by weeds out- competing the Arnica plants outweighs any positive effects on the sesquiterpene lactone yield in the flowers, then mechanisms of weed control should be applied.
In addition, whilst competition with weeds might increase the sesquiterpene lactone yield, depending on the harvesting method applied, the same weeds may contaminate the final product thereby reducing the value of the extract. As an added complication, if Arnica is surrounded by weeds that are more palatable to slugs, these may reduce herbivory damage to the flowers themselves, which could
235 in turn cause sesquiterpene lactone production not to be stimulated to the same extent as it would have been otherwise.
In order to investigate these factors further, the consequences of fertiliser and weeding treatments were investigated by monitoring dry weight yield, flower number, numbers of plants surviving, essential oil and extract quality. Treatments were combined in order for interaction effects to be determined.
4.2 Materials and Methods
4.2.1 The sites
4.2.1.1 AM3
In 2002, prior to the commencement of this study, a trial of A. montana (AM3), was established at the Agronomy Institute. In April 2003, a factorial combination of weed control and fertiliser treatments, each at two levels, was applied to the plots
(Table 4.1)
Table 4.1 - AM3 treatments
Code Treatment F +W+ Fertiliser and weeding F-W+ No fertiliser and weeding F+W- Fertiliser but no weeding F -W- No fertiliser and no weeding
The experimental design was a randomised complete block, with four replicates, each containing four plots, one for each experimental treatment. Experimental plots contained sixteen plants, of which an inner core of four were used for
236 harvesting, whilst the outer twelve ‘guard’ plants were intended to provide a buffer between the treatments. The trial was on a slight seaward slope and so the blocks were placed across the slope (Fig 4.1).
1 2 5 6 9 10 13 14 3 4 7 8 11 12 15 16
17 18 21 22 25 26 29 30 19 20 23 24 27 28 31 32
33 34 37 38 41 42 45 46 35 36 39 40 43 44 47 48
49 50 53 54 57 58 61 62 51 52 55 56 59 60 63 64
Figure 4.1 - Trial AM3
Each square represents a plant, each numbered square a study plant White = F-W-, light grey= F+W-, dark grey = F-W+ and black= F+W+.
In 2005 AM3 was abandoned due to the spread of disease ( Phytophthora and
Pythium DEFRA 2005), which reduced the number of repetitions of each treatment to below required levels (Figs 4.2 and 4.3).
237
Figure 4.2 - Spread of disease in AM3 (2005)
1 2 5 6 9 10 13 14 3 4 7 8 11 12 15 16
17 18 21 22 25 26 29 30 19 20 23 24 27 28 31 32
33 34 37 38 41 42 45 46 35 36 39 40 43 44 47 48
49 50 53 54 57 58 61 62 51 52 55 56 59 60 63 64
Figure 4.3 - Spread of disease in AM3 (2006)
Each square represents a plant, each numbered square a study plant. White = F-W-, light grey= F+W-, dark grey = F-W+ and black= F+W+, horizontal stripes = diseased or dead plants.
4.2.1.2 AM5
In 2005, due to the loss of plants in AM3, the trial was replaced by a new one
(AM5), on a block of A. montana planted in 2004. The same factorial combination
238 of fertiliser and weed control treatments (Table 4.1) was applied to AM5 as had been used for AM3 (Fig 4.4). Although this trial was designed to investigate the same factors, the number of replicates was increased to seven in order to improve statistical precision and to decrease the impact of reduced treatment replication in the event of further plant infection. In 2006, it was noted that plants in AM5 had also begun to suffer from crown rot (Fig. 4.5). However, it was felt that enough flowers would be available for extract to be produced, and so AM5 was not abandoned. However, yield was calculated as flower number and weight ‘per plant’ to compensate for the loss of plants.
4.2.2 Treatment regime
Both AM3 and AM5 had a light coating of dung when dug and Paraquat was sprayed around the outside of the plots in the first year in order to keep access weed free. Where the plants were deemed diseased (the leaves had wilted, turned yellow and only damaged or no flower stalks were present) they were excluded from the harvest. Fertiliser was applied only to plots marked F+ and application rates were equivalent to those applied to the yield plots (section 3.2.1). Nitrogen was applied as ammonium nitrate, phosphorous as phosphate and potash as potassium pentoxide, combined together in the form of slow release solid pellets
(N:P:K, 9:19:15). Application rates were equivalent to 40 kg/ha of N, 90.0 kg/ha of
P2O5, 67.4 kg/ha of K 2O. The pellets were applied in April at the start of each field season and care was taken to spread evenly over the treated plots and not into adjacent areas.
239
1 2 5 6 9 10 13 14 3 4 7 8 11 12 15 16
17 18 21 22 25 26 29 30 19 20 23 24 27 28 31 32
33 34 37 38 41 42 45 46 35 36 39 40 43 44 47 48
49 50 53 54 57 58 61 62 51 52 55 56 59 60 63 64
65 66 69 70 73 74 77 78 67 68 71 72 75 76 79 80
81 82 85 86 89 90 93 94 83 84 87 88 91 92 95 96
97 98 101 102 105 106 109 110 99 100 103 104 107 108 111 112
Figure 4.4 - Trial AM5 (2005)
Each square represents a plant, each numbered square a study plant. white = F-W-, light grey = F-W+, dark grey = F+W- and black = F+W+, horizontal stripes = dead plants
240
1 2 5 6 9 10 13 14 3 4 7 8 11 12 15 16
17 18 21 22 25 26 29 30 19 20 23 24 27 28 31 32
33 34 37 38 41 42 45 46 35 36 39 40 43 44 47 48
49 50 53 54 57 58 61 62 51 52 55 56 59 60 63 64
65 66 69 70 73 74 77 78 67 68 71 72 75 76 79 80
81 82 85 86 89 90 93 94 83 84 87 88 91 92 95 96
97 98 101 102 105 106 109 110 99 100 103 104 107 108 111 112
Figure 4.5 - Trial AM5 (2006)
White = F-W-, Light Grey = F-W+, Dark Grey = F+W-, Black = F+W+, horizontal stripes = diseased or dead plants. Although there were a number of fatalities, there were enough dried flowers to produce solvent extract from the majority of the plots
Only plots marked W+ and the border areas to the AM3 and AM5 plots were weeded. The first weeding in AM3 occurred on 28 th April 2004 and this continued twice weekly from 7 th June 2004 throughout the growing season. Weeding was done mostly by hand, with the occasional use of a hoe. In 2005, a wetter colder
241 spring (section 3.3.2), led to growth starting later and because of this the twice weekly weeding did not start until 5 th of May 2005.
A weeding treatment was applied to AM5 on 30 th June 2005 and weeding occurred twice weekly from this point on throughout the growing season. No flowers were harvested until 2006 in order to allow any treatment effect to become detectable in the flowers. The twice weekly weeding started again in AM5 on 10 th April 2006, continuing throughout the growing season as before. Unfortunately this was the last harvest, as disease spread quickly through this trial and by 2007 it was unusable and hence, abandoned.
It is important for a plant to quickly and efficiently occupy a large soil volume during establishment if it is to out perform rival plants (Dunbabin, 2007). The most prevalent weeds found on both AM3 and AM5 were buttercups ( Ranunculus acris
L.and R. repens L.) and dandelions ( Taraxacum officinalis F.H Wigg ) which also grew close to the base of the A. montana plants and were difficult to remove.
Other common weeds included: pineapple weed ( Chamomilla suaveolens Pursh), rosebay willowherb ( Chamerion angustifolium L.), broadleaf dock ( Rumex obtusifolius L.), daisies ( Bellis perennis L), yorkshire fog ( Holcus lanatus L.) and wavy hair-grass ( D. flexuosa L. ).
4.2.3 Harvesting
Flowers were harvested as per section 3.2.3 between 14:00 and 16:00 hours each day. Flowers from AM3 were stored in labelled brown paper bags at c.a. 4°C,
242 while those from AM5 were ground and stored in labelled centrifuge tubes under nitrogen at c.a -20°C.
4.2.4 Extractions
Flowers were harvested from AM3 in the summers of 2004 and 2005. Towards the end of 2004 work by Douglas et al. (2004) suggested that the ‘over’ (section 3.1.3) flowers may contain more secondary compounds than the partially open flowers and so the over flowers from the border plants of the AM3 plots were collected.
Although it could be claimed that any detectable effect would be less reliable as there were essentially ‘no borders’ for these flowers, it was hoped that the increased sample size would compensate for the reduced number of repetitions that would otherwise have been available.
Due to the large amount of flower material required for essential oil extraction, only the ‘normal’ and ‘over’ flowers from 2004 were subjected to hydrodistillation as described in section 2.2.2.1. The flowers from AM3 in 2005 and AM5 in 2006 were extracted by the solvent method as described in section 2.2.2.2.
4.2.5 Analysis
Data on flower dry weight, flower number, essential oil content, essential oil components and sesquiterpene lactone content were analysed to determine whether there was a relationship between these aspects of yield and fertiliser/weeding treatment. Samples extracted were analysed as per sections
2.2.3, 2.2.4 and 2.2.5.
243 A 2-way ANOVA was used to test for significance and the data were blocked by row. A 2-way ANOVA assumes the data are sampled from a Gaussian or normal distribution. To check this, the data was plotted for all analyses. Data that were not normally distributed were transformed and where this was done the type of transformation has been described. The residual errors were also plotted to ensure that they were random and independent of treatment. Genstat was used for all analysis.
A two way ANOVA (blocked by row) was run to determine whether there was any link between plant survival and treatment. Blocking removes a degree of freedom for each level, and although this technique may remove some sensitivity from the analysis, it can be an effective way of reducing the residual error. The trials in this research were on a seaward slope and as this could have an effect on drainage, blocking is one way of taking this into account.
As described earlier, disease claimed plants in both the AM3 and AM5 trials and eventually led to both being abandoned. It was important to establish whether this disease was related to the treatment in any way as this could affect how yield should be calculated. If disease is unrelated to treatment, then diseased plants should be excluded and yield calculated on a per plant basis. However, if disease is related to treatment, then yield needs to be calculated per plot in order to take this corresponding loss of yield into account.
Disease led to some missing data which can imbalance a 2 way blocked ANOVA.
To adjust for this, when the Genstat encounters a missing value in a table, it
244 automatically estimates the missing value on the basis of the treatment and block means. It then performs an ordinary analysis of variance as though the estimates were genuine variations and removes a degree of freedom from the error term for each missing term estimated. This can lead to problems when there are a considerable number of missing values (e.g. AM3) and so this was taken into account when interpreting the results.
4.3 Results
4.3.1 Plant survival
The number of sample plants (border plants excluded), were recorded in 2004,
2005 and 2006 for AM3 (Figure 4.6) and 2005 and 2006 for AM5 (Figure 4.7).
4
3.5
3
2.5 Number of plants 2 surviving
1.5
1
0.5
0 F-W- 2004 F+W+ 2005 F+W-
Year 2006 F-W+ Treatment
Figure 4.6 - Average number of plants per plot surviving in each trial by treatment (AM3)
245
4
3.5
3
2.5
Number of plants 2 surviving
1.5
1
0.5
0
F-W- 2005 F+W+
2006 F+W-
Year 2007 F-W+ Treatment
Figure 4.7 - Average number of plants per plot surviving in each trial by treatment (AM5)
Due to the high number of zero readings for the individual plots and the ordinal nature of the data, the data were transformed (log+10) before running a two way
ANOVA on the results. It was found that the null hypothesis, that treatment had no effect on plant survival, could not be rejected for log transformed AM3, nor for untransformed AM3 blocked by row. However, the residual variation was high for all the tests, which may be due to the pattern of disease spread (i.e. from one corner inwards as figure 4.3 suggests). Ideally it would have been beneficial to have had the option of blocking by column, but this was not possible due to the design of the trial and repetitions of treatment being by row (to compensate for the slope) and not by column.
246 However, AM5 had more repetitions of each treatment and although when survival was scored for 2006 no evidence of any effect of treatment was found, in 2007 it was found that the null hypothesis could be rejected for weeding treatment for log transformed AM5 data (blocked by row) using a critical alpha of 0.01 (F( 1,18 ) = 8.49 with p = 0.009), but not for fertiliser or interaction (Fig. 4.6). The data were transformed due to the high death rate among the plants, but the null hypothesis could also be rejected for non transformed AM5 data albeit with less confidence
(critical alpha of 0.05 (F( 1,24 ) = 7.95 with p = 0.011). The results strongly indicate that A. montana was more likely to die or suffer disease in 2007 if it had been weeded (Fig 4.7). It should be noted however, that in 2006 this effect was not evident and so this could either mean that the treatment takes some time to become apparent, or that in late 2006 or early 2007, weeding was particularly aggressive and caused ‘one off’ damage. Regardless, as the null hypothesis could not be rejected for the 2006 results, the number of plants in each plot should still be taken into account when calculating yield.
247
1.07
1.06 Not weeded (W-) Weeded (W+)
1.05
plants surviving (log+10) surviving plants
1.04
Number Number of
1.03
No Fertiliser (F-) Fertiliser (F+) Fertiliser
Figure 4.8 - Log transformed plant survival data for AM5 in 2007, where the treatment is fertiliser at different levels of weeding
Bar represents standard error of differences of means
4.3.2 Yield
4.3.2.1 AM3
No effect of treatment or interaction was found for total flower dry weight, number of flowers, or number of flowers per plant or for average flower dry weight for AM3 in 2004 (Fig 4.9 and Table 4.2). In addition, there was no significant effect of treatment on average flower dry weight in 2004 (Fig 4.10). The same lack of treatment effect was found for the over flowers gathered from AM3 in 2005 (table
4.3). Although in all cases the background variation was very high, it should be noted that no fertiliser and no weeding gave the lowest flower yield.
248 80.00 400.00
70.00 350.00
60.00 300.00
50.00 250.00
40.00 200.00 Weight Weight (g) 30.00 150.00 Number of flowers
20.00 100.00
10.00 50.00
0.00 0.00 F+W+ F-W+ F+W- F-W- Treatment
Total flower dry weight Flower No Flowers per plant
Figure 4.9 - Yield data for AM3 in 2004
F+ = fertiliser added, F- = no fertiliser, W+ = weeded and W- = no weeding treatment. Bars represent std deviation.
0.25
0.20
0.15
0.10
Average flower dry flower weight Average
0.05
0.00 Ave F+W+ Ave F-W+ Ave F+W- Ave F-W-
Figure 4.10 - Average flower dry weight for AM3 in 2004
F+ = fertiliser added, F- = no fertiliser, W+ = weeded and W- = no weeding treatment. Bars represent standard deviation.
249
Table 4.2 - Yield data for AM3 (2004)
Average total Average Average total Ave number Treat- flower dry Ave flower dry number of flower number of plants per ment weight per plot weight (g) flowers per per plot plot (g) plant F+W+ 213.75 37.91 0.18 4 53.44
F-W+ 248.25 49.48 0.20 4 62.06
F+W- 243.75 38.93 0.17 4 60.94
F-W- 158.25 29.58 0.19 4 39.56 SED 55.0 8.61 0.017 0 13.74 (5%) F/W F/W F/W F/W F/W F/W F/W F/W F/W W W W W W W W W W W F F F F F F F F F F p valuesp
0.524 0.455 0.149 0.858 0.147 0.112 0.773 0.652 0.967 0.524 0.445 0.149 / / /
F+ = fertiliser added, F- = no fertiliser, W+ = weeded and W- = no weeding treatment
Table 4.3 - Yield data for the over flowers in AM3 (2005)
Average total Average total Average number Treat- Ave flower dry Ave number of flower number flower dry weight of flowers per ment weight (g) plants per plot per plot per plot (g) plant
F+W+ 13.75 2.57 0.19 3.00 4.58
F-W+ 21.25 4.63 0.21 3.50 5.96
F+W- 21.00 3.71 0.17 3.00 6.97
F-W- 10.00 1.68 0.16 3.50 2.69
SED 9.25 2.028 0.02597 1.00 2.482 (5%) F/W F/W F/W F/W F/W F/W F/W F/W F/W F/W W W W W W W W W W W F F F F F F F F F F p valuesp 0.794 0.786 0.185 0.993 0.542 0.180 0.808 0.136 0.350 0.283 0.930 0.425 0.807 0.135 1.0
F+ = fertiliser added, F- = no fertiliser, W+ = weeded and W- = no weeding treatment
4.3.2.2 AM5
As described in section 4.3.1, the data on plant survival indicated that yield for
AM5 in 2006 should be calculated on a ‘per plot’ basis Taking this into account, in
250 2006 for AM5 (Table 4.4), no effect of treatment on average flower weight, total flower number or total dry weight was found.
There were indications of a slight effect of fertiliser on the number of flowers per plant (Fig 4.11), at a critical alpha of 0.10 (F( 1,18 )= 3.10, p=0.095), and for flower dry weight per plant (Fig 4.12) (F( 1,18 )=4.05, p= 0.059). In both cases, the application of fertiliser was weakly associated with increased yield: increased flower production leading to greater dry weight when diseased or dead plants were excluded from the analysis. However, the background variation was considerable.
Table 4.4 Yield data for AM5 (2006)
Average Average total Average Average Total flower Average number of Treat- flower dry weight number of dry weight flower sample ment number per per plant, flowers per (g) weight (g) plants per plot per plot (g) plant per plot plot F+W+ 9.52 56.29 0.17 3.29 2.99 17.58
F+W- 5.89 34.00 0.19 2.86 2.22 12.75
F-W+ 6.52 39.57 0.17 2.86 1.85 11.25
F-W- 4.50 28.57 0.17 3.00 1.31 8.26 SED 2.90 17.67 0.020 1.00 0.823 5.13 (5%) F/W F/W F/W F/W F/W F/W F/W F/W F/W F/W F/W F/W W W W W W W W W W W W W F F F F F F F F F F F F p p values 0.295 0.181 0.699 0.384 0.195 0.656 0.244 0.623 0.433 0.725 0.725 0.484 0.091 0.268 0.846 0.149 0.292 0.801
251 18
16
Not weeded (W-) 14 Weeded (W+)
12
Number of flowersplant per 10
8
No Fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.11 - Number of flowers produced per plant by fertiliser treatment at different levels of weeding, AM5 (2006)
Bar represents standard error of differences of means
3.00
2.75
2.50 Not weeded (W-) Weeded (W+)
2.25
2.00
1.75 Dry weight per(g) plant
1.50
1.25 No fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.12 – Flower dry weight produced per plant by fertiliser treatment at different levels of weeding, AM5 (2006)
Bar represents standard error of differences of means
252
4.3.3 Essential oil
No effect of treatment on the amount of essential oil produced per gram of flower material was found for AM3 in 2004 (Table 4.5). However, as stated earlier, the number of repetitions of the treatments was very low and this may be masking any effect. As described in Chapter 3, A. montana essential oil is solid at room temperature and this often led to difficulties in the process of extraction. This combined with the very low levels of oil obtained in the first place led to the loss of six samples prior to analysis and to a further reduction in degrees of freedom for the analysis (3). However, sample loss only occurred once for the over flowers that were collected from the border plants which led to a greater number of degrees of freedom (12) for the analysis.
Table 4.5 - Yield of oil by treatment for AM3
Treatment Yield ml/g F+W+ 0.0015 F-W+ 0.0012 F+W- 0.0014 F-W- 0.0015 SED (5%) 0.0091 F/W F/W W W F F
p values 0.574 0.574 0.574
Although no effect of treatment on amounts of oil produced was detected for either partially open or over flowers, the profile of the two oils was significantly different
(Table 4.6) with 30 compounds in the partially open flowers compared to 20 in the
253 over flowers. There was some evidence of a significant effect of fertiliser on both heptacosane (critical alpha 0.05 F (1,3) =24.24, p=0.016) (Fig 4.13), and peak D
F(1,3) =16.88, p=0.026) (Fig 4.14) in the partially open flowers, with considerably more of both present if fertiliser was not added. There was also an effect of weeding on perhydrofarnesylactone (critical alpha 0.05 F (1,9) =11.34, p=0.043) with more present in oil from non-weeded plants (Fig 4.15). However, the very large number of missing samples in this case is likely to have weakened the reliability of the Genstat analysis.
For the essential oil extract from over flowers, there was no evidence of a significant effect of fertiliser or weeding on levels of essential oil compounds.
1.0
0.8
Not weeded (W-) Weeded (W+) 0.6
0.4 Percentage content
0.2
0.0
No fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.13 – Percentage content of heptacosane in the essential oil of partially open flowers. by fertiliser treatment at different levels of weeding, AM3 (2004)
Bar represents standard error of differences of means
254
Table 4.6 - Profile of essential oil from partially open and over flowers
Partially open flowers Over Flowers Compound F+W+ F-W+ F+W- F-W- F+W+ F-W+ F+W- F-W-
Camphene 0.29 0.06 0.17 0.00 0.00 0.00 0.00 0.00 Phellandrene
Cymene
Linalool 0.16 0.00 0.27 0.00 0.08 0.26 0.05 0.28 Nonanal 0.09 0.00 0.16 0.00 0.00 0.00 0.00 0.06
Decanal
Thymol 0.27 0.14 0.15 0.00 0.00 0.00 0.00 0.00 Caryophyllene <(E)> 30.86 38.36 38.44 38.08 7.76 11.74 6.85 11.45
Bergamotene
Tetracosane
255
0.8
0.6
Not weeded (W-) Weeded (W+) 0.4
0.2
Percentage content 0.0
-0.2
No Fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.14 – Percentage content of peak D in the essential oil of partially open flowers by fertiliser treatment at different levels of weeding, AM3 (2004)
Bar represents standard error of differences of means
5.75
5.50
5.25 Not weeded (W-) Weeded (W+)
5.00
4.75 Percentage content 4.50
4.25
No fertiliser (F-) Fertiliser (F+) Fertiliser
Figure 4.15 – Percentage content of perhydrofarnesylactone in the essential oil of partially open flowers by fertiliser treatment at different levels of weeding, AM3 (2004)
Bar represents standard error of differences of means
256
4.3.4 Sesquiterpene lactone production
In 2004 all the flower material from AM3 was used for oil extractions. Although some flowers were produced in a number of plots in AM3 in 2005, there was not sufficient to run repetitions for all the treatments, so this trial was abandoned.
However, enough flower material was produced for solvent extraction in all but three of the plots in AM5 in 2006. The three low yield plots were excluded from the analysis and three degrees of freedom removed from all of the 2-way blocked
ANOVAs in this section (F (1,15) ). The results of the analysis are shown in Table 4.7.
The results for total content and helenalin and dihydrohelenalin esters are shown in Figure 4.16. The variation was high, but a significant interaction effect was detected for isobutyryl helenalin and treatment (critical alpha 0.05 F (1,15) = 5.57, p=0.032) (Fig 4.17) in that higher amounts of this compound were produced in plots that had fertiliser added and were weeded, but in the absence of fertiliser, a better yield was produced when plants were not weeded. A similar relationship was also found for isovaleroyl helenalin (critical alpha 0.05 F (1,15) = 5.13, p=0.039)
(critical alpha 0.05 F (1,15) = 4.72, p=0.046) with more produced overall in plots where fertiliser was added (Fig 4.18). There was another interaction effect detectable for tigloyl dihydrohelenalin (critical alpha 0.05 F (1,15) = 5.29, p=0.036)
(Fig 4.19) but the levels of this compound were often below the official LOQ.
The aforementioned interaction effect was almost detectable for total helenalin esters (critical alpha 0.10 F (1,15) = 4.19, p=0.059) (Fig 4.20) and total sesquiterpene lactones (critical alpha 0.05 F (1,15) = 4.64, p=0.048) (Fig 4.21). There was also an effect of fertiliser on total helenalin ester content (critical alpha 0.05 F (1,15) = 4.70,
257 p=0.047) and total sesquiterpene lactone content (critical alpha 0. 05 F (1,15) = 4.62, p=0.048).
Table 4.7 – ANOVA results for significance table of variation in sesquiterpene lactones by weeding and fertiliser treatment in AM5 (2005)
Treatment Fertiliser Weeding Interaction
F p F p F p
Dihydrohelenalin / / / / / /
Helenalin 1.90 0.189 0.51 0.488 0.32 0.580
Acetyl helenalin 0.40 0.539 0.43 0.524 1.94 0.184
Acetyl dihydrohelenalin 2.79 0.116 0.02 0.883 0.01 0.933
Isobutyryl helenalin 2.05 0.173 0.51 0.488 5.57 0.032
Isobutyryl dihydrohelenalin 0.01 0.940 0.31 0.583 3.18 0.095
Methacryl helenalin 0.56 0.467 0.34 0.569 2.89 0.110
Methacryl dihydrohelenalin 0.53 0.478 1.44 0.249 1.78 0.202
Isovaleroyl helenalin 4.72 0.046 0.10 0.755 5.13 0.039
2-methylbutyryl helenalin 3.23 0.093 0.37 0.552 4.01 0.064
Isovaleroyl dihydrohelenalin 0.17 0.687 1.62 0.222 3.11 0.098
2-methylbutyryl dihydrohelenalin 0.73 0.407 0.15 0.701 2.09 0.169
Tigloyl helenalin 2.05 0.173 0.69 0.421 1.53 0.235
Tigloyl dihydrohelenalin 2.67 0.123 0.53 0.477 5.29 0.036
Total Helenalins 4.70 0.047 0.14 0.714 4.19 0.059
Total dihydrohelenalins 1.10 0.310 0.64 0.437 4.23 0.058
Total sesquiterpene lactones 4.62 0.048 0.20 0.664 4.64 0.048
Significant (F crit = 4.54, p<0.05) Not significant but above 4.0 Not significant
258 6.00 F+W+ F+W- F-W+ F-W-
5.00
4.00
3.00
Content(mg/ml) 2.00
1.00
0.00 Total dihydrohelenalins Total helenalins Total sesquiterpene Total quantifiable lactones sesquiterpene lactones
Figure 4.16 – Content of sesquiterpene lactones in the solvent extract of AM5 (2006) under fertiliser (F+/F-) and Weeding (W+/W-) treatments
Bars represent standard deviation
1.00
0.95
0.90
Not weeded Weeded 0.85
0.80
Content (mg/ml) 0.75
0.70
0.65
0.60 No Fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.17 – Content of isobutyryl helenalin in the solvent extract of AM5 (2006) by fertiliser treatment under different levels of weeding
Bar represents standard error of differences of means
259
0.75
0.70
0.65
Not weeded Weeded 0.60
0.55
Content (mg/ml) 0.50
0.45
0.40
No fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.18 – Content of isovaleroyl helenalin in the solvent extract of AM5 (2006) by fertiliser treatment under different levels of weeding
Bar represents standard error of differences of means
0.035
0.030 Not weeded Weeded
0.025
(mg/ml) Content 0.020
0.015
No Fertiliser (F -) Fertiliser (F+) Fertiliser
Figure 4.19 – Content of tigloyl dihydrohelenalin in the solvent extract of AM5 (2006) by fertiliser treatment under different levels of weeding
Bar represents standard error of differences of means
260
3.8
3.6
Not weeded Weeded 3.4
3.2
Content (mg/ml)
3.0
2.8
No fertiliser (F-) Fertiliser (F+) Fertiliser
Figure 4.20 – Total helenalin ester content (mg/ml) in the solvent extract of AM5 (2006) by fertiliser treatment under different levels of weeding
Bar represents standard error of differences of means
4.0
3.8
Not weeded Weeded 3.6
3.4
Content (mg/ml) 3.2
3.0
2.8 No fertiliser (F-) Fertiliser (F+) Fertiliser
Figure 4.21 – Total sesquiterpene lactone content (mg/ml) in the solvent extract of AM5 (2006) by fertiliser treatment under different levels of weeding
Bar represents standard error of differences of means
261 4.4 Discussion
As outlined in the previous chapter, the harvest index of a crop can vary considerably, with nutrient availability and weeds both playing an important role.
Whilst increases in the former normally have a positive correlation with yield and the latter a more negative relationship, crops that require either fertiliser or weeding are both more expensive to grow (particularly if hand weeding is required) and susceptible to damage either during the process of fertiliser addition, or while the weeds are being removed. Weeds can also serve as a reservoir of crop pests, particularly where the weed’s growing season is longer than that of the crop (Tivy,
1990). They can also indirectly decrease the leaf area and photosynthetic efficiency via shading.
The weeds identified in Orkney tended to be shorter than the stems of the arnica flowers and so would be unlikely to affect the quality of the flower harvest via contamination, even if mechanical harvesting were to be employed. However, the more similar the weeds are to the growth habit of the crop the more potential damage the weeds can cause (Forbes & Watson, 1992). If Arnica does not establish control over local root space quickly, it is likely that this space will be taken by rival plants. Dandelions can quickly out-compete the rosette of A. montana because they are hard to distinguish from A.montana until they are already established (personal observation) . Crops such as A. montana are more susceptible to infestation by weeds as due to their growth habit there is a long space of time in which there are wide open spaces, unlike A. chamissonis where the growth habit leads to faster ground cover. Weeds are a recurrent and nearly universal threat to crop productivity whether it is for the crop in question or for their
262 potential to spread and damage the yield of surrounding crops (Liebman &
Gallandt, 1997) and so their management is vital. However, it is also important that the impact of any weed removal or fertiliser treatment on both the quantity and quality of the flowers is measured, particularly when the production of the active components may be related to stress and/or defence.
No effect of weeding on plant survival was apparent neither in the AM3 trial, nor for AM5 in 2006. For the former, it is possible that any link between weeding and plant death was masked by the small number of replicates, while for AM5 it could be argued that the treatments may not have been fully established by 2006. There was evidence for a correlation between weeding and plant death in 2007, which would be in contrast to the findings of Pegtel (1994), who determined that a light disturbance of the topsoil layer (weeding, turf cutting etc) favoured the growth of A. montana by enhancing its competitiveness against grasses. On the other hand, the soil found in Orkney is heavier than that of the aforementioned Pegtal study, and it may be that weeding in such heavy soil causes physical damage to the surrounding A. montana plants. It is possible that the plants are more exposed to damage by factors such as wind when weeded, but the lack of effect in any other year combined with the much earlier first weeding session in 2007 (occurring when the A. montana plants were still quite small), may make root damage from the movement of heavy soil more likely. Regardless, because the number of surviving plants in AM3 and AM5 in 2005 and 2006 appeared to be unrelated to treatment, the number of plants in each plot was taken into account when the effect of treatment on total dry weight and flowers per plant was calculated. This did reduce
263 some of the residual variation and may have made other effects of treatment easier to detect.
No effect of weeding or fertiliser treatment on yield was found for AM3 in 2004 though there was some indication that the addition of fertiliser increased both number of flowers and flower dry weight per plant for AM5 in 2006, with most produced in the fertilised and weeded plots. No effect was found on total or average flower dry weight, which suggests that the addition of fertiliser may support an increased number of flowers (although the effect was minimal), but has no effect on flower size.
Pegtel, (1994), found that the addition of fertiliser to the soil surface layer
(mimicking the effects of ammonia deposition) caused A. montana to quickly be out-competed by Deschampsia sp. He stressed the importance of maintaining an open vegetation structure at a low level of nutrient supply, which could either be created through competition for nutrient resource or by a low starting level in the first place. Smallfield & Douglas (2008) found that A. montana responded to moderate levels of fertiliser, but that high levels in spring encouraged crown rot.
However, for both AM3 and AM5, the control of no fertiliser and no weeding (F-W-) had the lowest flower yields.
In contrast to the findings of Azizi et al. (2008) and de la Fuente et al. (2003) neither fertiliser nor weeding seemed to have an impact on essential oil yield.
There was some evidence that the application of fertiliser influenced the content of the oil, with increased levels of heptacosane and peak D in AM3, while higher
264 amounts of perhydrofarnesylactone were found in non weeded plots. However, the very low degree of freedom for this analysis, and the need for a high number of
Genstat substitutions make this result highly unreliable. There was also no agreement between these results and those obtained from over flowers which provides more evidence that the significance result may have been an artefact of the low sample size. There did seem to be higher amounts of caryophyllene in the essential oil from the over flowers in fertilised plots. Caryophyllene is believed to have anti-inflammatory properties (Chaven et al. 2010) which may mean that the quality of A. montana extract would be improved by fertiliser application if it is to be grown for essential oil extract.
There also seemed to be a relationship between the application of fertiliser and increased isovaleroyl helenalin, total helenalin ester and total sesquiterpene lactone content with more produced in plots where fertiliser was added. The highest levels of sesquiterpene lactones overall were to be found in extracts from plots that both had fertiliser added and were weeded. However, there was an interaction effect detectable where, in the absence of fertiliser, a better yield was produced when the plots were not weeded. The effect was strongest for isobutyryl helenalin and isovaleroyl helenalin, as well as for total helenalins and total sesquiterpene lactones.
This could be because when the plants were fertilised and there were no other plants in the immediate vicinity, the plant could ‘afford’ to invest more in the production of defensive compounds. Conversely, adding fertiliser to plots with weeds present, increases the competitiveness of the rival plants to the extent that
265 the A. montana plants may have to divert resources away from the production of defence compounds to growth. It is also possible that plants which are shaded by weeds have less exposure to sun, and as described in section 3.3.6, this study found an association between lower levels of sesquiterpene lactone production and a low number of total sun hours over the year (section 3.3.6).
Helenalin esters are more effective defence compounds than the dihydrohelenalin forms due to the α-methylene-γ-lactone and cyclopentenone electrophilic centre
(Wagner et al. 2004b). They have been found to cause the greatest depletion in glutathione in insects (Guillet et al. 2000) and so could be of greater benefit in flower defence. The elevated levels found in this study would hence suggest an increased capability for defence.
Although some effect of treatment on dihydrohelenalin esters was detected, mostly the levels of these compounds were so low that they were almost not detectable. It is possible that the reason this effect is more apparent in the helenalin esters than the dihydrohelenalin esters could be because more helenalin was produced and hence the results were more reliable and quantifiable. The trend for these compounds did seem to be opposite to that of the helenalin esters though, and it may be that there is a trade off in their production, or that this is another indicator of the existence of a hydrogenase system converting the sesquiterpene lactones between the helenalin to the dihydrohelenalin type as suggested by Schmidt et al.
(1997).
266 In agronomic terms, obviously a balance needs to be found between reducing weeds, increasing yield and minimising costs. Both fertiliser and weeding seem to increase yield, but hand weeding is an expensive and potentially damaging process (AM5 2007). Not weeding could save costs, and not fertilising would remove the potential advantage to weeds, but the plots which were not weeded and did not have fertiliser added had the lowest flower dry weight and flower number production of all the plots. The most straightforward approach would be to chemically control the weeds, but flowers produced for herbal preparations are normally required to be organic and so this is not possible. Hand weeding is acceptable but the labour required would be expensive and would reduce crop viability. Hence the research programme went on to investigate alternative methods of weed control and the effects that these have on yield and plant survival.
267 5. The influence of planting regime on yield
5.1 Introduction
The previous trials have illustrated some of the ways in which environmental conditions and agronomic treatments may influence the yield of Arnica in Orkney.
However, correlation does not guarantee causation and it may be that complex interrelationships are clouding the causative factor. For example, high rainfall appears to be correlated with a reduction in yield (section 3.3.5), but it may be that this is due to an association of rain with another factor, such as decreased total hours of sun.
Whilst it would have been interesting to have studied each environmental factor separately in order to have determined which was most influential, it would have been both beyond the scope of this study and of limited practical use in the field.
Instead, trials of practicable agronomic regimes that can help limit the effects of high rainfall such as water-logging and poor drainage were undertaken, in order to determine whether they could help mitigate the negative effects of climate on flower yield and extract quality.
For A. montana there appears to be a correlation between weeding treatment and high plant mortality, while the absence of weeds is correlated with an improved quality of floral extract (section 4.3.4). The high plant mortality in 2007 (section
4.3.1) was possibly caused by root damage during hand weeding, which might have resulted in the breakage of the roots when the surrounding heavy soil was disturbed. The subsequent root damage may then have created potential infection
268 sites for disease. It could be that this infection then triggered an increased production of sesquiterpene lactones, with the effect strongest in plots that also had fertiliser added. Although no significant effect of treatment was found on flower yield, there was some indication that more flowers were produced under fertiliser treatment.
One of the aims of this study was to determine the effect of agronomic conditions on crop yield and quality and to determine the ‘ideal’ regime for Arnica grown in
Orkney. Results from the fertiliser and weeding trials, indicate that this regime should be one that either prevents weeds from becoming established, or one in which it is possible to remove the weeds without damaging the roots and stems of the Arnica plants.
Whilst for the majority of commercial crops this would be a selective herbicide, this is not an option for a crop that is predominantly sold as organic (Burnie, unpublished). An alternative to herbicide would be the application of mulch, which serves as a barrier to rival plants and helps prevent them from becoming established. Although mulch can result in a decrease in yield due to a decreased density of plants per hectare, the yields of the conventional plots in flat beds also decreased over time, as outlined in section 3.3.3. If the use of mulch stabilises plant numbers, then this may more than make up for the decrease in yield caused by the application of a mulching treatment.
Galambosi et al. (1998) found that A. montana grew well under black plastic mulch in Finland. In New Zealand trials, Smallfield and Douglas (2008) tested a variety
269 of organic weed control methods including a fatty acid herbicide, wood chip mulch, corn starch film, paper mat material and synthetic woven cloth. The organic, fatty acid herbicide trialled was deemed to be not effective and as the Arnica didn’t establish fast enough, both the paper and the corn starch film failed to last the season. High winds (section 3.3.2) would make the latter treatments difficult to apply in Orkney. The New Zealand trials found that a synthetic mat was effective at controlling weeds, whilst the weed mat and the wood mulch both produced larger plants. In Orkney, the most readily available mulch materials in these categories were sawdust (organic) and polythene (synthetic) as well as quarry fines (mineral), and so trials into the effectiveness of all three against a control of no mulch were conducted.
In trials of the effect of organic and mineral mulches on soil, Iles & Dosmann
(1999) found that soil temperatures were highest under mineral mulch and in control plots, while soil moisture was highest under organic mulch. The authors attributed this to the organic mulches meshing together more effectively than the coarser textured mulches, although under cool wet conditions the difference was not significant. Readings of pH are normally lowered under organic mulches and shredded bark because of the leaching of organic acids from the decomposition of plant material (Himelick & Watson, 1990 and Hild & Morgan, 1993). They can be raised in some cases (Iles & Dosmann, 1999) though it is likely in such cases that there has been a temporary elevation of ammonium ions leached from the mulch which will be gradually removed via nitrification. Investigations into the exact changes to the soil caused by mulching treatments were beyond the scope of this study, but if any of the mulches were found to be effective, the mechanisms
270 outlined above might provide indications of other, potentially effective, agronomic treatments.
Although mulches can reduce the effects of competition with weeds, they can also limit water evaporation from soil (Davies et al. 2001) and when soil is saturated with water plant roots can become damaged due to the associated poor aeration and accumulation of toxins. Normally, damage only occurs when the soil is at 5°C or above and the root tissue is respiring, so waterlogging in winter rarely causes damage (Davies et al. 2001). However, due to the Oceanic climate in Orkney, the soil is regularly above this minimum and so high rainfall may be more likely to lead to root damage. For this reason, as well as the association of crown rot with waterlogging (DEFRA, 2005), agronomic regimes that combined mulching with increased drainage were also investigated.
Ridges can improve drainage (Davies et al. 2001) and as A. montana is believed to prefer free draining soil (Pegtel, 1994) it may be that planting A. montana in such beds will lead to an improvement in yield. Smallfield and Douglas (2008) in their New Zealand trials found that the highest yields were obtained on heavy soil when ridges were used, though flat planting was best for light and free draining soils. The soil in Orkney is heavy (section 3.3.1) and so a trial comparing the yields of Arnica planted in ridges with those grown on the flat was conducted.
No data comparing the growth of A. chamissonis on ridges to those under mulch were available. This species has been shown to be resistant to both crown rot and weeds (section 3.3.3) and so it could be concluded that such trials would be
271 unlikely to be required. However, the Orkney trials have also demonstrated that the extract of A. chamissonis flowers contains far fewer sesquiterpene lactones than that of A. montana and of those found in the literature. A. chamissonis is found more frequently in wet soil (Cronquist, 1955) such as that found in Orkney. It may be that the growth of A. chamissonis in more free draining conditions would be a stress on the plant. Whilst in conventional agriculture this practice is normally to be avoided, it is possible that exposing the plant to increased levels of stress will stimulate the production of an increased number and/or levels of sesquiterpene lactones. Such a change in composition would make the extract more comparable to that of A. montana and hence would make the extract more valuable.
In order to investigate these hypotheses further, trials of mulching and ridge treatment were conducted for both A. montana and A. chamissonis . For A. montana the aim was to investigate whether the treatments would increase the physical yield and/or quality of extract, whilst for A. chamissonis it was to establish whether increasing the stresses these plants were exposed to would stimulate production of sesquiterpene lactones and hence increase the value of the extract.
Of course, if none of the treatments has an effect on either species, it may be that another factor, such as provenance, has more of an influence on yield than the environment.
It has been known for some time that infraspecific variation in the production of secondary metabolites can exist between plants which are morphologically identical (Lawrence, 2000). Such production is under strict regulation in plant cells,
272 co-ordinated and controlled by transcription factors which in turn can be triggered and regulated by a number of factors (Endt et al. 2002). These factors can be internal or external signals which then lead to a controlled response. As the environments in which plants grow can be highly variable, differentiated taxa emerge via the action of natural selection on existing variations in the population.
Such control is exerted by the plant’s DNA and is hence inheritable (Lawrence,
1974).
Selective breeding techniques can be used to artificially create ‘ideal’ plants which respond to environmental signals in the desired way. For such techniques to be successful however, the extent to which environment controls the factors of interest needs to be determined. For example, in a study of populations of A. montana in the Ukranian Carpathian mountains Kobiv (1993) found that the physical characteristics of this species were limited by territory and monotype.
They found no specific relation to ecological or topological factors; concluding that leaf and stem characteristics were inherited. However, Kahmen & Poschlod (2000) in their principle component analysis of the fitness traits of A. montana , concluded that phenotype was more influenced by the environment than by genetic variation.
Bomme and Daniel in their 1994 ARBO trial, demonstrated that selective breeding could result in increased yields of A. montana flowers. In a subsequent study, it was determined that levels of sesquiterpene lactones in this variety did not vary with altitude in plots between 590 and 2230m (Spitaler et al. 2006). This would seem to give weight to the hypothesis that genome - and in its more practical sense, seed source - is a strong influencer of yield in A.montana . Another
273 breeding programme has commenced in New Zealand (Douglas et al. 2004), though the results are not yet available.
As described earlier (section 1.8.3.3), it has been established that there are two main chemotypes of A. montana : a Spanish type that contains mostly dihydrohelenalin esters and a central European type that contains mainly helenalin esters (Willuhn et al. 1994). However, recent research into variation within the
Spanish chemotypes has revealed that A. montana flowers harvested from heathland areas at high altitude had a similar helenalin content to those harvested in central Europe, whilst the extract of flowers harvested from Spanish peat bogs and meadow environments had higher levels of dihydrohelenalin (Perry et al.
2009). This is in contrast to the work by Spitaler et al. (2006) who found similar levels of sesquiterpene lactones at all altitudes in transplanted A. montana of the
ARBO variety. If sesquiterpene lactone production in A.montana is genetically controlled as this research suggests, it may be that seed from different sources will also produce variable levels of sesquiterpene lactones in the Orkney environment.
In a study of North American A. chamissonis, Willuhn et al. (2004) identified five chemotypes from different locations on the west coast (Figure 2.1 and table 2.3).
Recordings of such variation in natural populations gave weight to the belief that sesquiterpene lactone production in A. chamissonis is naturally more variable than that of A. montana (Willuhn et al. 2004) . The A. chamissonis grown in this study would appear to consistently match that of ‘type E’ (mostly chamissonolide) and so far has shown very little variation both in terms of quality of extract and in consistency of yield .
274
It is possible that the A. chamissonis seed purchased for this study was originally sourced from one of the ‘E locations’ and it is this that has led to the consistent profile. However, A. chamissonis is found in a wider range of environments than A. montana and it may be that the environment, not chemotype, has the strongest influence on yield. If the local conditions exert more control over yield for this species than they do for A. montana , then a different chemotype grown in identical conditions, would produce a similar sesquiterpene lactone profile. To this end, a trial of both A. montana and A. chamissonis plants produced from a variety of seed sources was conducted.
5.2 Materials and Methods
5.2.1 Trial sites and planting materials
All plots had a light coating of dung when dug, with Paraquat sprayed around the outside of the plots in the first year in order to keep access weed free. All plots were on a seaward facing slope and were established at the Agronomy Institute in
Kirkwall, Orkney.
In March prior to the commencement of each trial, seeds of A. montana (in 2005 for AM7, in 2006 for ACM1 and ACM2) and A. chamissonis (in 2006 for ACM1 and
ACM2) were added to light, peat free compost and kept moist and in polythene tunnels till germination. All plants were moved out of the polythene tunnels in May for hardening, before being transplanted into the marked plots. Where the plants were deemed unhealthy (the leaves had wilted or turned yellow) they were
275 discarded. After planting, when disease became apparent (yellow, wilted leaves and damaged or stunted flower stalks), these plants were excluded from analysis.
5.2.1.1 Mulch (AM7)
Plot AM7 was planted in June 2005 in order to create a trial that would determine which type of mulch would be most effective at controlling weeds. Seeds were purchased from CN Seeds and germinated as per section 5.2.1.
The trial used a random block design with four mulch treatments: quarry dust
(mineral), sawdust (organic), polythene (synthetic) and a control (no mulch), with seven replicates by row (Figs 5.1 and 5.2). It was felt that for any successful treatment to be economically viable, the mulches used in this trial should be readily available in the local area. To this end the sawdust was sourced from a local mill, the quarry dust from a local quarry and the polythene sheeting was commercially available. The trial was on a slight seaward slope and so the blocks were placed across the slope. Experimental plots contained twenty plants of which an inner core of six were used for harvesting, while the outer fourteen ‘guard’ plants were intended to provide a buffer between the treatments. The plot was covered with polythene with either planting holes cut out (synthetic mulch treatment) or large squares cut out (organic, mineral and control treatments). The synthetic cover was cut open as the plants grew and a path was left between plant rows for ease of harvesting and weeding access. Samples of the mineral mulch were sent to Celtest Ltd (Bangor) for analysis.
276 The synthetic mulch was easiest to apply, and only had to be applied once as long as the holes were expanded in size as the plants grew. The sawdust mulch was relatively easy to apply, although messy (particularly when windy), and was the cheapest of the four options being a waste product from a local sawmill. The mineral mulch was heaviest and hardest to transport, but was again a local waste product and was believed locally to be a soil improver (Martin, personal communication).
1 2 3 7 8 9 13 14 15 19 20 21 4 5 6 10 11 12 16 17 18 22 23 24
25 26 27 31 32 33 37 38 39 43 44 45 28 29 30 34 35 36 40 41 42 46 47 48
49 50 51 55 56 57 61 62 63 67 68 69 52 53 54 58 59 60 64 65 66 70 71 72
73 74 75 79 80 81 85 86 87 91 92 93
76 77 78 82 83 84 88 89 90 94 95 96
97 98 99 103 104 105 109 110 111 115 116 117 100 101 102 106 107 108 112 113 114 118 119 120
121 122 123 127 128 129 133 134 135 139 140 141 124 125 126 130 131 132 136 137 138 142 143 144
145 146 147 151 152 153 157 158 159 163 164 165 148 149 150 154 155 156 160 161 162 166 167 168
Figure 5.1 – AM7
White = No mulch, Light Grey = Sawdust mulch, Dark Grey = Quarry fines and Black = Polythene
277
Figure 5.2 – Photograph of AM7 (2005)
Although fertiliser is often required for mulched beds (Himelick & Watson, 1990) it was felt that this would not be required during the first year of the study.
5.2.1.2 Ridges and mulch (ACM1)
Plant mortality was high in AM7 and so a further trial of mulching treatment
(ACM1) was created in June 2006 (Figs 5.3 and 5.5). Preliminary observations of
AM7 had suggested that organic mulch was the most effective and so this treatment was included within a factorial design of mulch, ridge and species. The plots were 1m x 10cm (Fig 5.4) with 0.8 m between rows. As the creation of ridges required more space which would reduce the yield per m 2, planting was relatively close, with seedlings planted at 12.5cm intervals. With 216 plants in 18.4m x 24 metres, the density was just over 8 plants per metre of bed. Guard rows were placed on either side of every treatment row and mulch was reapplied at the start of each field season. Plots were weeded as necessary, which was normally every week during the growing season.
278
Figure 5.3 – Photograph of ACM1 (2006)
Guard row
12.5cm
Guard row
Figure 5.4 – Typical plot of ACM1
279 0.8m 0.8m
Guard Guard Guard 0 Guard Guard 0 0 Guard 0 Guard Guard . . . . 8 8 8 8 m m m m
AM AM AM AM AM AM AC AC AC AC AC AC AC AM AM AM AM AM AM AC AC AC AC AC AC
AM AM AM AC AC AC AC AC AC AC AC AC
AM AM AM AM AM AM AM AM AM AC AC AC AC
AM AM AM AM AM AM AM AM AM AC AC AC
AM AM AM AC AC AC AC AC AC AC AC AC
AM AM AM AC AC AC AC AC AC AC AC AC
AM AM AM AM AM AM AM AM AM AC AC AC
AM AM AM AC AC AC AC AC AC AC AC AC AC
AM AM AM AM AM AM AM AM AM AC AC AC
AM AM AM AM AM AM AC AC AC AC AC AC
AM AM AM AM AM AM AC AC AC AC AC AC
Column 1 2 3 4
Figure 5.5 – ACM1
Green = border plant, yellow = sawdust, heavy lines = raised, thin lines = flat, AC = A. chamissonis and AM = A. montana
280 5.2.1.3 Ridges and seed source (ACM2)
In order to establish the extent to which the source of seed influenced the production of sesquiterpene lactones, plot ACM2 was created in June 2006 (Fig
5.6 & 5.7). A. montana seeds sourced from CN Seeds (CNS) and Jelitto (JEL)
(German/East European wild harvested) were germinated alongside seeds of the cultivar ARBO donated by Saatzucht Steinach GmbH (ARB) (Bomme & Daniel,
1994) and seeds from untreated Orkney plots harvested in 2005 (ORK). A. chamissonis seeds were purchased from Horizon Seeds (HOR), Van Duesen
(VAN), Jelitto (JEL) and CN Seeds (CN). Although most suppliers could not identify their seed source, it was felt that by using a range of companies to source the seed it was highly likely the seed came from a range of source locations.
Both A. montana and A. chamissonis were planted in ridges in an attempt to ensure maximum drainage. The plots were 1m x 10cm with 0.8 m between rows and one row of eight plants in each plot (as per Fig 5.4). A guard row was planted on either side of the trial and with 384 plants in 6 x 8 metres the plant density was just over 8 plants/m of bed. Neither mulch nor fertiliser was added.
Figure 5.6 – Photograph of ACM2 (2006)
281
AC AM AM AM AM AC AC AC CNS CNS ARB ORK JEL VAN JEL HOR
AM AC AC AM AM AC AM AC JEL VAN JEL CNS ARB HOR ORK CNS
AC AC AM AM AM AM AC AC VAN CNS CNS ARB ORK JEL HOR JEL
AM AC AC AC AM AM AC AM ARB JEL HOR CNS CNS ORK VAN JEL
AM AM AC AC AC AC AM AM CNS JEL VAN JEL HOR CNS ARB ORK
AM AM AM AC AC AC AM AC ORK ARB JEL HOR CNS JEL CNS VAN
Figure 5.7– ACM2
AC = A. chamissonis and AM = A. montana. CN Seeds (CNS), Jelitto (JEL), Horizon Seeds (HOR), Van Duesen (VAN), ARBO (ARB) and Orkney (ORK).
5.2.2 Chemical analysis
Flowers from each plot in all trials were harvested as described in section 3.2.3 and were weighed and dried at 40°C till constant we ight (normally 4 days). After this, flowers from each plot were combined, ground and stored in labelled centrifuge tubes under nitrogen at c.a -20°C.They w ere extracted, identified and quantified as described in sections 2.2.2.2, 2.2.3 and 2.2.5 respectively.
5.2.3 Statistical analysis
Genstat was used for all statistical analysis. For the data from AM7 a one way
ANOVA blocked by row was performed, followed by Dunnett’s test, in order to
282 compare the results of each treatment to the control (no mulching treatment). A one way ANOVA was also conducted for ACM2, but as Dunnett’s test would have been inappropriate in this case (there was essentially no control), Tukey-Kramer’s test was applied instead. The latter allows variable populations sizes to be taken into account, which was felt to be important considering the high number of plant fatalities in previous trials. For ACM1 a two way ANOVA blocked by row was conducted.
For both ACM1 and ACM2, analyses were run separately for A. montana and A. chamissonis . The two species are so different in terms of physical yield, plant number and sesquiterpene lactones produced; a combined analysis would be meaningless.
5.3 Results
5.3.1 Plant survival
5.3.1.1 AM7
Prior to the commencement of the AM7 trial, samples of the quarry fines were sent for analysis against ICRCL (Inter-Departmental Committee on the Redevelopment of Contaminated Land) standards and by X-Ray fluorescence (XRF), in case mineral leaching might be a potential problem. The results are displayed in tables
5.1 and 5.2 but there were no potential problems identified, with levels of all parameters being well within the recommended margins. The pH was slightly high for A. montana though, as it normally prefers poor acidic soils (Maguire, 1943).
283 Table 5.1 – ICRCL analysis of the mineral mulch (Celtest, 2004)
Mineral mulch Parameter (mg/ml unless stated) Arsenic <10 Cadmium <1 Lead <50 Mercury <50 Selenium <1 Water soluble boron <1000 Copper <20 Nickel <50 Iron 410 Zinc <100 Electrical Conductivity (µScm) 94 Chemical Oxygen Demand (COD) 41 Polycyclic Aromatic Hydrocarbons (PAHs) <0.2 Phenols <0.5 Complex cyanide <0.05 Free cyanide <0.05 Ammonia <0.1 Total Sulphate (SO 4) 24 Sulphide <10 pH (unit) 8.9 Chloride <10
Table 5.2 - XRF analysis of the mineral mulch (Celtest, 2004)
Component Result (%) Fe (total) 3.36 CaO 12.45 SiO 2 45.11 MgO 5.21 Al 2O3 11.01 P 0.035 P2O5 0.080 Mn 0.056 MnO 0.072 S 0.260 SO 3 0.649 K2O 3.080 V2O5 0.015 TiO 2 0.560 BaO 0.060 ZnO 0.010 Na 2O 1.800 Cr 2O3 0.030 LOI (Loss on ignition) 15.18
284 In October 2005 mortalities were already evident in AM7 (Fig 5.8) and by 2006 there were very few plants remaining (Figs 5.9, 5.10 and Table 5.3). When the results were analysed by treatment and blocked by row, it was found that there was a significant effect of mulching treatment on log transformed plant survival at a critical alpha of 0.05 (F 1,18 =3.94, p=0.025). A significantly lower number of plant mortalities occurred when plants were grown under the sawdust mulch (Fig 5.11)
(when the same test was conducted on non log transformed data the difference was not significant). The biggest difference in plant survival was between the polythene and the sawdust treatment, whilst neither the polythene mulch nor the quarry fines were significantly different from the control.
Although none of the mulching treatments were particularly effective at preventing weeds from becoming established (Fig 5.12), on the sawdust mulch the weeds were easier to remove. This seemed to be because the weeds became established on the thick layer of sawdust, rather than going down into the soil and so the roots did not get a chance to embed themselves amongst the roots of the
Arnica plants. Under the polythene mulch the only point at which weeds could become established was right next to the Arnica plant. This made them particularly hard to remove and it is possible the process of removing them led to damage to the roots of the Arnica plants and subsequent infection and crown rot. This perhaps explains the higher number of plant deaths in plots under the polythene mulch treatment compared to those grown under the sawdust mulch.
This trial was abandoned in the summer of 2006 due to the high number of fatalities and the most effective mulch (sawdust) was applied to the ACM1 trial.
285
1 2 3 7 8 9 13 14 15 19 20 21 4 5 6 10 11 12 16 17 18 22 23 24
25 26 27 31 32 33 37 38 39 43 44 45 28 29 30 34 35 36 40 41 42 46 47 48
49 50 51 55 56 57 61 62 63 67 68 69 52 53 54 58 59 60 64 65 66 70 71 72
73 74 75 79 80 81 85 86 87 91 92 93
76 77 78 82 83 84 88 89 90 94 95 96
97 98 99 103 104 105 109 110 111 115 116 117 100 101 102 106 107 108 112 113 114 118 119 120
121 122 123 127 128 129 133 134 135 139 140 141 124 125 126 130 131 132 136 137 138 142 143 144
145 146 147 151 152 153 157 158 159 163 164 165 148 149 150 154 155 156 160 161 162 166 167 168
Figure 5.8 – AM7 in October, 2005
White = No mulch, Light Grey = Sawdust mulch, Dark Grey = Quarry fines and black = Polythene Striped = diseased or dead plants
Figure 5.9 – AM7 in 2006
286
1 2 3 7 8 9 13 14 15 19 20 21 4 5 6 10 11 12 16 17 18 22 23 24
25 26 27 31 32 33 37 38 39 43 44 45 28 29 30 34 35 36 40 41 42 46 47 48
49 50 51 55 56 57 61 62 63 67 68 69 52 53 54 58 59 60 64 65 66 70 71 72
73 74 75 79 80 81 85 86 87 91 92 93
76 77 78 82 83 84 88 89 90 94 95 96
97 98 99 103 104 105 109 110 111 115 116 117 100 101 102 106 107 108 112 113 114 118 119 120
121 122 123 127 128 129 133 134 135 139 140 141 124 125 126 130 131 132 136 137 138 142 143 144
145 146 147 151 152 153 157 158 159 163 164 165 148 149 150 154 155 156 160 161 162 166 167 168
Figure 5.10 – AM7 in June 2006
White = No mulch, Light Grey = Sawdust mulch, Dark Grey = Quarry fines and black = Polythene Striped = diseased or dead plants
Table 5.3 – Number of plants surviving in AM7 (2006)
Average Mulching number of Treatment surviving plants Control 1.86 Sawdust 4.00 Quarry fines 1.86 Polythene 1.14
Values in plants per plot - maximum possible was 4
287
1.14
1.12
1.10
Log plantLog survival 1.08
1.06
No mulch Organic Mineral Synthetic Mulching treatment
Figure 5.11 – Number of plants surviving, by mulching treatment
Bar represents standard error of differences of means
Figure 5.12 – Photograph illustrating the lack of effectiveness of the mulches against chickweed (Oct 2005)
288 5.3.1.2 ACM1
There was a significant effect of bed type on the survival of A. montana in 2007
(Fig 5.13) when blocked by column at a critical alpha of 0.05 (F1,17 = 7.29, p=0.015), indicating that more A. montana plants died in flat beds than ridges, regardless of whether mulch was used. There was some indication (though not significant) that mulch may have slightly increased plant mortalities when it was applied to a flat bed. It should be noted however, that the apparent increased survival of A. montana when planted in ridges was only apparent when blocked by column, and not when blocked by block as was the original experimental design.
Although this potentially weakens the conclusions that can be drawn from this analysis, as shown in Table 5.4, there were much higher mortalities in column 4 than in any other column. It is possible that crown rot had spread from adjacent plots and increased plant mortalities on the side of column 4. The effect was also most marked in the flat beds (personal observation). It was concluded that altering the analysis to block by column, whilst altering the original design of the analysis, was vital to compensate for the very clear spread of disease.
Table 5.4 – Number of plants surviving in ACM1 by column
Column Number of plants surviving 2006 2007 Total per column 48.00 45.00 1 Average per plot 8.00 7.50 Total per column 48.00 48.00 2 Average per plot 8.00 8.00 Total per column 47.00 46.00 3 Average per plot 7.83 7.67 Total per column 34.00 12.00 4 Average per plot 5.67 2.00
289
1.23
1.22
Mulch No mulch 1.21
1.20
1.19 Log of Log number of plants surviving
1.18
Raised Flat Type of bed
Figure 5.13 – Log of number of plants surviving against bed at different levels of mulch
Bar represents standard error of differences of means
Unfortunately, no effect of treatment on the survival of A. chamissonis could be investigated, because it was not possible to count individual plants due to the rapid production and establishment of new shoots. This made it impossible to determine the exact number of A. chamissonis plants in each plot. Certainly, if it were not for each of the plots being marked out with string, it would have been hard to distinguish between each plot of A. chamissonis. If there were an effect of treatment, it is likely it would have been relatively minor and more likely to be detectable in measurements of flower production (section 5.3.3).
5.3.1.3 ACM2
There was no significant effect of seed source on plant survival for A. montana in
2007 either when blocked by row or unblocked. Three plots were lost to disease,
290 but this did not impact significantly on the trial as each plot was from a different seed source (Orkney, Jelitto and CN Seeds). There was also no detectable effect of seed source on survival for A. chamissonis although, as for ACM1, this was hard to measure exactly due to the difficulty in determining exact plant numbers for this species.
5.3.2 Flower production
5.3.2.1 AM7
AM7 was established in 2005 which meant it did not flower until 2006.
Unfortunately, as described in section 5.3.2, by 2006 there were already a high number of deaths. This in turn affected the number of replicates in each treatment.
No significant effect was detected for mulching treatment either on number of flowers produced or on fresh and dry flower weight, but the low number of treatment replicates may have masked any significant effect of treatment on yield.
5.3.2.2 ACM1
There was a highly significant effect of mulch on average dry weight (Fig 5.14) and average fresh weight of A. montana flowers at a critical alpha of 0.001
(F 1,13 =17.57, p=0.001 and F 1,13 =8.13, p=0.014, respectively) when blocked by column. The plants grown under mulch produced flowers of a higher average dry weight than those not grown under mulch. However, there was no effect of treatment detected in number of flowers or total dry weight produced. This was unexpected, as normally an increase in flower weight would lead either to an increased total flower weight with a similar number of flowers, or a similar total weight with a decreased total number of flowers. When dry weight and number of
291 flowers were calculated on a per plant basis, there was no detectable effect of either treatment.
For A. chamissonis , a significant effect of mulching treatment was found for both the total dry weight (Fig 5.15) and total fresh weight produced (Fig 5.16) at critical factor 0.05 (F 1,20 = 12.27, p=0.002 and F 1,20 =12.14, p=0.002 respectively) with more flower material produced under mulch treatment than without. When blocked by column, the result was more significant but only marginally (F 1,17 = 13.99, p=0.002 and F 1,17 = 14.02, p=0.002 respectively). More flowers were produced per plot under mulch at critical factor 0.001 (F 1,20 =14.58, p=0.001) when not blocked by column (Fig 5.17) and (F 1,17 =15.95, p<.001) when blocked by column. The reduced effect of blocking by column for A. chamissonis compared to A. montana is likely due to the former having a lower susceptibility to crown rot .
0.076
0.074
Mulch 0.072 No mulch
0.070
0.068
Averagedryweight (g)
0.066
0.064 Raised Flat Type of bed
Figure 5.14 – Average dry weight of A. montana flowers against bed at different levels of mulch
Bar represents standard error of differences of means
292
11
10 With mulch
No mulch 9
Dry weight (g) 8
7
6
Raised Flat Type of bed
Figure 5.15 – Total dry weight of A. chamissonis flowers against bed at different levels of mulch
Bar represents standard error of differences of means
55
50
Mulch 45 No mulch
40
Fresh weight (g)
35
30
Raised Flat Type of bed
Figure 5.16 – Total fresh weight of A. chamissonis flowers per plot against bed at different levels of mulch
Bar represents standard error of differences of means
293
180
160 Mulch No mulch
140
flowers of Number 120
100
Raised Flat Type of bed
Figure 5.17 – Total number of flowers of A. chamissonis flowers per plot against bed at different levels of mulch
Bar represents standard error of differences of means
5.3.2.3 ACM2
No effect of seed source was found for average dry weight, average fresh weight, total dry weight, total fresh weight, total number of flowers or number of flowers produced per plant for A. montana in ACM2. Similarly, no effect of seed source on average dry weight, average fresh weight, total dry weight, total fresh weight or number of flowers produced, by treatment, was detectable for A. chamissonis in the same trial.
5.3.3 Sesquiterpene lactone production
5.3.3.1 AM7
The high number of plant mortalities in AM7 and the low number of flowers produced by the remaining plants, led to insufficient plant material being produced
294 for solvent extraction in 16 of the 28 plots. This reduced the number of replications to the extent that for extract analysis of the flowers, only one plot was available from the mineral mulch treatment, two for the control, three for the synthetic and four for the organic treatment. Probably because of this, no significant effect of treatment was found on production of any of the sesquiterpene lactones.
5.3.3.2 ACM1 A. montana
Although in ACM1 there were a number of A. montana plots in which there were either no flowers, or not enough flower material for analysis, significant effects of treatments on some of the individual sesquiterpene lactones were apparent (Table
5.5). For example, when blocked by column, acetyl helenalin seemed to be produced to the greatest extent in treatments with no mulch and in ridges (Fig
5.18) at a critical factor of 0.05 (F 1,12 = 6.68, p=0.024 and F1,12 = 4.99, p=0.045 respectively), although this was the only sesquiterpene lactone for which this joint relationship was evident. This would theoretically be the driest of the treatments, as the lack of mulch would mean there was no impedance to evaporation, and the ridge would have the most effective drainage.
Isobutyryl dihydrohelenalin was found in higher concentrations in plots with mulch treatment applied (F 1,12 = 4.77, p=0.049), regardless of the type of bed (Fig 5.19), whilst methacyrl helenalin was found in higher amounts in plots which had no mulching treatment (F 1,12 = 8.14, p=0.015) (Fig 5.20). It should be noted though, that isobutyryl dihydrohelenalin was found in very low amounts, mostly below the
LOQ, and this may be an artefact of background interference. Tigloyl helenalin on the other hand, displayed a relatively strong interaction effect (Fig 5.21) with
295 higher amounts found under no mulch on a ridge, and under mulch on the flat
(F 1,12 = 5.98, p=0.031).
Table 5.5 – Table of F-values for sesquiterpene lactones by treatment in ACM ( A. montana)
Treatment Bed Mulch Interaction F p F p F p Dihydrohelenalin 1.69 0.217 3.10 0.104 0.53 0.483
Helenalin 3.81 0.075 0.16 0.692 0.46 0.512
Acetyl helenalin 6.70 0.024 5.01 0.045 3.45 0.088
Acetyl dihydrohelenalin 2.98 0.110 1.29 0.278 2.04 0.179
Isobutyryl helenalin 0.82 0.382 0.07 0.798 0.40 0.537 Isobutyryl 0.16 0.700 4.75 0.050 0.11 0.743 dihydrohelenalin Methacryl helenalin 2.59 0.133 8.16 0.014 1.52 0.242 Methacryl 0.45 0.517 0.02 0.879 0.09 0.771 dihydrohelenalin Isovaleroyl helenalin 0.01 0.914 0.04 0.849 1.67 0.221
2-methylbutyryl helenalin 0.41 0.536 0.03 0.866 0.84 0.377 Isovaleroyl 0.71 0.417 1.70 0.217 2.68 0.128 dihydrohelenalin 2-methylbutyryl 0.35 0.563 4.60 0.053 0.78 0.395 dihydrohelenalin Tigloyl helenalin 0.24 0.632 1.10 0.314 6.00 0.031
Tigloyl dihydrohelenalin 2.64 0.130 4.27 0.061 0.12 0.737
Total Helenalins 2.27 0.158 2.11 0.172 1.24 0.287
Total dihydrohelenalins 0.04 0.849 0.65 0.437 0.90 0.363 Total sesquiterpene 2.16 0.168 1.84 0.200 1.30 0.277 lactones
Significant (f crit =4.75, p<0.05) Not significant but above 3.5 Not significant
296
1.2
1.0
Mulch No mulch 0.8
0.6 Concentration(mg/ml)
0.4
Raised Flat Type of bed
Figure 5.18 – Acetyl helenalin in A. montana extract against bed at different levels of mulch
Bar represents standard error of differences of means
0.065
0.060
Mulch g/ml) 0.055 No mulch
0.050
Concentration(m 0.045
0.040
0.035 Raised Flat Type of bed
Figure 5.19 – Isobutyryl dihydrohelenalin in A. montana extract against bed at different levels of mulch
Bar represents standard error of differences of means
297
1.6
Mulch 1.4 No mulch
1.2
Concentration (mg/ml)
1.0
Raised Flat Type of bed
Figure 5.20 – Methacyrl helenalin in A. montana extract against bed at different levels of mulch
Bar represents standard error of differences of means
0.60
0.55
0.50 Mulch
No mulch 0.45
0.40 Concentration (mg/ml)
0.35
0.30
Raised Flat Type of bed
Figure 5.21 – Tigloyl helenalin in A. montana extract against bed at different levels of mulch
Bar represents standard error of differences of means
298
5.3.3.3 ACM1 A. chamissonis
In contrast to the results found for A. montana, when blocked by column no significant effect of the treatments on any of the levels of sesquiterpene lactones in
A. chamissonis extract was detectable (Table 5.6).
Table 5.6 – Table of F-values for sesquiterpene lactones by treatment in ACM1 (A. chamissonis)
Treatment Bed Mulch Interaction F p F p F p Helenalin 0.01 0.907 0.45 0.510 0.45 0.510 4-O-acetyl-6- 0.75 0.400 0.63 0.439 0.16 0.697 desoxychamissonolide Chamissonolide 2.20 0.156 0.96 0.341 0.22 0.648
Arnifolin 2.19 0.157 1.01 0.330 0.20 0.659
Total Helenalins 2.26 0.151 1.04 0.322 0.23 0.638
Total sesquiterpene lactones 2.25 0.152 1.04 0.322 0.22 0.643
5.3.3.4 ACM2 A. montana
As for ACM1, in the ACM2 trial there were also a number of A. montana plots in which there were either no flowers, or not enough flower material for analysis.
However, a clear effect of seed source on concentration of sesquiterpene lactones was evident (Figs 5.22, 5.23 and Table 5.7) with the highest values of most of the individual sesquiterpene lactones occurring in the floral extract of the CN Seed sourced plant material. Plants grown from seeds from Jelitto also produced chamissonolide which no other A.montana plants grown in this environment produced.
299 Amount (mg/ml) 10.00 12.00 14.00 16.00 18.00 0.00 2.00 4.00 6.00 8.00
lactones5.22 in Figure –Sesquiterpene
eeaisDhdoeeais Total sesquiterpene lac Dihydrohelenalins Helenalins r neyOrk
ARBO
Bars represent standard deviation deviation standard represent Bars
montana A.
CNSeeds extract
(ACM2) sources seed differentfrom
Jelitto
tones
300 Table 5.7 – Table of F-values for sesquiterpene lactones by treatment in ACM2 (A. montana)
Average (mg/ml) ANOVA result Treatment CN Orkney Arbo Jelitto F p Seeds
Dihydrohelenalin 0.01 bc 0.00 b 0.03 ac 0.01 bc 13.48 0.004
Helenalin 0.03 0.04 0.06 0.05 5.03 0.045
Acetyl helenalin 1.28 bc 0.86 b 1.80 ac 0.95 bc 6.10 0.030
Acetyldihydrohelenalin 0.08 b 0.07 b 0.23 a 0.11 b 31.45 <.001
Isobutyryl helenalin 1.69 b 1.81 b 4.21 a 1.36b 24.04 <.001
Isobutyryl dihydrohelenalin 0.05 b 0.05 b 0.31 a 0.05 b 56.28 <.001
Methacryl helenalin 1.28 1.10 1.77 1.49 2.07 0.205
Methacryl dihydrohelenalin 0.07 bc 0.06 b 0.23 ac 0.11 bc 6.18 0.029
2-methylbutyryl helenalin 1.23 c 1.41 c 2.20 a 1.15 bc 11.53 0.007
Isovaleroyl helenalin 0.75 0.48 0.88 0.50 2.97 0.119 2-methylbutyryl 0.02 0.02 0.07 0.04 11.72 0.006 dihydrohelenalin c bc a bc Isovaleroyl 0.02 0.03 0.10 0.04 7.15 0.021 dihydrohelenalin b bc ad cd Tigloyl helenalin 0.47 0.58 0.57 0.72 1.27 0.367
Tigloyl dihydrohelenalin 0.02 b 0.03 b 0.08 a 0.04 b 21.08 0.001
Chamissonolide 0.00 0.00 0.00 0.23 a 5.66 0.035
Total Helenalins 6.74 b 6.27 b 11.48 a 6.45 b 18.67 0.002
Total dihydrohelenalins 0.26 b 0.26 b 1.04 a 0.40 b 25.41 <.001 Total sesquiterpene 7.00 6.53 12.52 6.85 24.38 <.001 lactones b b a b
Means with different subscripts are significantly different from each other (p<0.05, bold values indicate p<0.01) by the Tukey-Kramer test.
301
Amount (mg/ml) 0.00 1.00 2.00 3.00 4.00 5.00 6.00 D ihydrohelenalin
sesquiter of individual Figure5.23 – Concentration H e lenali n Acetyl
hel e na A lin cet y ldihy d rohelenali
n Isob Orkney u tyrylhelenalin
Is ob uty ryldihydrohelenal
in Meth Bars represent standard deviation deviation standard represent Bars a cr
Sesquiterpene lactoneSesquiterpene yl ARBO Methacryl helenalin sources(ACM2) dihy d rohel e pene lactones in in pene lactones 2- nal methyl inB
butyrylhelenal
CN Seeds Is in
ovaleroyl helen
2 -methylb
A. montana A. u a ty lin ryldihydrohelenalin
Isovale
royl d ih y
extract dr Jelitto oh ele n alin
Tigloylhelenalin
from different from seed
T igloy ldihy d rohelenal
in C hamissonol
ide
302
5.3.3.5 ACM2 A. chamissonis
In contrast to the results found for A. montana, when blocked by column no significant effect of the treatments on the levels of sesquiterpene lactones in A. chamissonis extract was found (Table 5.8 and Fig 5.24).
Table 5.8 – Table of F-values for sesquiterpene lactones by treatment in ACM1 (A. chamissonis)
Seed Average (mg/ml) source Treatment CN Orkney Arbo Jelitto F p Seeds Helenalin 0.02 0.01 0.02 0.02 0.59 0.632 4-O-acetyl-6- 0.22 0.24 0.22 0.19 2.14 0.141 desoxychamissonolide Chamissonolide 2.40 2.54 2.39 2.24 0.92 0.456
Arnifolin 0.39 0.28 0.37 0.37 2.77 0.081
Total helenalins 3.03 3.08 3.00 2.81 0.43 0.733 Total sesquiterpene 3.03 3.08 3.00 2.81 0.43 0.733 lactones
303
Amount (mg/ml) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50
Figure 5.24 – Sesquiterpene lactones5.24 in Figure – Sesquiterpene Average Horizon
4-O-Acetyl-6- Helenalin
Average Van Duesen desoxy
Bars indicate standard deviations deviations standard indicate Bars
lactone Sesquiterpene A. chamissonisA. hmsooieAnflnTotalSLs Arnifolin Chamissonolide
Average CNSeeds(AC)
extract
(ACM2) different from sources seed
Average Jelitto(AC)
304 5.4 Discussion
In Chapter 4 it was determined that weeding may be correlated with an increased risk of plant mortality, but that the reduced competition may have led to increased yields of sesquiterpene lactones. It was this that led to the trials of organic, mineral and synthetic mulch.
None of the mulching treatments were particularly effective at reducing weed numbers, although the weeds on the sawdust mulch were much easier to remove.
It is likely that it was this that led to reduced plant mortality under the sawdust treatment, rather than any reduction of competition for resources. Under the synthetic mulch, where the weeds had become established, they were harder to remove than in the controls, as they were by nature of the treatment, forced to grow closer to the base of the A. montana plants. This is believed to have contributed to the particularly high mortality in these plots, in contrast to the suggestions by Pegtel (1994) that light disturbance of the topsoil layer, such as weeding, favours the growth of A. montana by enhancing its competitiveness against grasses.
With reduced levels of evaporation under the plastic mulch, it is possible that excess water may have contributed to the spread of disease and the high number of mortalities. Waggoner et al. (1960) found that plastic mulch led to roots being attacked more often by the pathogenic Rhizoctonia solani . Kühn. Regardless of mechanism however, both the use of polythene and quarry fine mulches can safely be discarded as possible agronomic regimes for A. montana in Orkney, as
305 both were indistinguishable from the control. The sawdust mulch on the other hand seemed to lead to greater plant survival, and was hence applied to the ACM1 trial.
If the spread of crown rot is linked to waterlogging in the soil, methods of controlling levels of standing water are required for this crop to be successful in the Orkney environment. Whilst levels of rainfall cannot be controlled, agronomic regimes that allow for improved drainage can be implemented. It was to this end that the ridge treatment was applied to the ACM trial in addition to the mulching treatment, not least as the latter can exacerbate waterlogging by reducing surface evaporation.
Pegtel (1994) found that A. montana preferred free draining soil and the results of the ACM1 trial would seem to support this statement. More A. montana plants died when grown on flat than on ridges, which occurred regardless of whether mulch was used. A slightly higher mortality rate occured when the plants were grown on flat beds with mulch which was possibly due to the combination of reduced water drainage and reduced evaporation. On the flat, the problem of waterlogging could be exacerbated by mulch, which would lead to increased mortalities.
Smallfield & Douglas (2008) found that A. montana produced a higher yield when grown on ridged beds in heavy soil. However, although the soil in Orkney is relatively heavy (section 3.3.1), there was no increase in flower number, average flower weight or total dry weight of A. montana from this treatment.
306 A considerable amount of the residual variation in the analyses of plant survival was accounted for by blocking by column. This indicates that there was a source of variation that was influenced by column and had an effect on the number of plant fatalities. The most likely source of this variation was crown rot, spreading in from the adjacent, abandoned A. montana plots. Whilst an environmental factor such as wind damage or exposure to sun could be a factor, this would also have affected A. chamissonis. Instead, a much reduced effect of column was found for
A. chamissonis and so crown rot (to which A. chamissonis seems less susceptible) would be the most likely source of this residual variation.
No effect of either ridge treatment or mulch was found on survival of A. chamissonis plants, though it is possible this was masked due to the difficulties experienced in attempts to establish the exact number of plants present. An effect was found however, of mulching treatment on both total fresh and total dry flower weight, with a greater mass of both produced from plots grown under mulch. A greater number of flowers were also produced which indicates that the flowers that were produced were no larger than those of other plots, just that there were more of them. As plant number wasn’t known, this could indicate either an increase in plant number or that more flowers were produced by existing plants.
In the previous chapter it was found that extract from weeded A. montana plots, contained the highest levels of active compounds. This was not found in the mulched plots, other than an increase in acetyl and methacryl helenalin along with an interaction effect in tigloyl helenalin. This gives some weight to the theory that it was the damage to the roots that stimulated the increase in sesquiterpene lactone
307 production, not that lack of competition for resources. On the other hand, both average dry and average fresh flower weight increased in plots under mulch, and so there was no dilution effect and overall yield was still increased. An increase in plant size was found by Smallfield and Douglas (2008) in their trials of woodchip mulch, and this increase in flower size may be an indicator of larger plants, though they were not measured in this study. Decoteau et al. (1988) found a light coloured mulch reflected more blue light (which influenced the stem thickening and length of the tomato plants in the study) as well as PAR, than the black coloured mulch, and it may be this that has lead to increased flower yield.
Whilst there was no effect of treatment on the levels of the active compounds in A. chamissonis , again there was no indication that the increased flower dry weight produced by the mulched plants, diluted the levels of sesquiterpene lactones contained within them as levels were similar to that of the control.
In ACM2 there was a very clear effect of seed source on the levels of the active compounds in the floral extract of A. montana. The highest levels of most of the individual sesquiterpene lactones were from the plants grown from seed purchased from CN Seeds. Almost double the levels of helenalin and total sesquiterpene lactones, and almost three times the dihydrohelenalins were found in the extract from CN Seed material than that from the other sources. There was some variation in the levels of sesquiterpene lactones in the Jelitto, ARBO and
Orkney plants, but none of these were significant. The ARBO source, bred to produce a higher flower head yield (2-3 times more and 2-3 times bigger), than conventional ecotypes (Eickmeyer, 2005), produced the lowest yield of
308 sesquiterpene lactones and produced the lowest flower dry weight and number of all the plots (though the difference between them was not significant). This is perhaps not surprising as it was bred for cultivation in a high altitude environment, not for an island environment. What was surprising was the similarity between
ARBO and both the commercially available Jelitto material, as well as the Orkney produced seed. No effect of seed source on plant survival, flower number, flower weight or dry weight was found .
The above results provide some evidence for the existence of a number of chemotypes of A. montana. As described earlier, Perry et al. (2009) found that flowers harvested from high altitude sites produced higher levels of sesquiterpene lactones, whilst Spitaler et al. (2006) found that when the ARBO cultivar was planted at varying altitude, levels of sesquiterpene lactones produced were relatively unaffected. The highest levels of helenalin esters and lowest levels of dihydrohelenalin esters in the Perry et al. (2009) study, came from the three highest altitude sites (1330-1460m above sea level) and were all from heathland sites. There was however, no significant effect of altitude on sesquiterpene lactone content or composition below these altitudes.
These findings together suggest that not only is the production of sesquiterpene lactones under quite strict genetic control, but that more than two chemotypes of
A. montana exist. If the latter is the case, then there is a selective pressure (in
Spain at least) for having higher amounts of one type of sesquiterpene lactone over another, and that this factor may be a factor correlated with altitude (e.g. populations of T.arnicae increases with altitude, Scheidel et al. 2003).
309
The heathland samples had high levels of isobutyryl helenalin which was characteristic of the central European chemotype, but also high levels of methacryl, tigloyl and isovaleroyl dihydrohelenalins typical of the commercially available Spanish material. This suggests either two chemotypes have bred to create a mixed chemotype or that the sites contained individuals of both chemotypes. The authors described the heathland plants displaying characters closest to the subspecies montana , whilst the lower altitude plants had characters similar to that of atlantica (narrow basal leaves). They call for further research into the taxonomic status of Spanish A. montana ssp atlantica which was described by de Bolos y Vayreda (1947) to differ mainly in its reduced size. However, they also report a number of plants that ranged between ‘intermediate and indistinguishable’ to the A. montana found in the Alps.
No significant differences in sesquiterpene lactone production were found for any of the A. chamissonis extracts, regardless of seed source. It may be that all four of the seed sources came from the same original location, or came from plants with the same chemotype. It would account for the lack of variation in both sesquiterpene lactones and flower yield by treatment, but it may also be that an environmental factor such as total sun hours, is exerting more of an influence on the yield of A. chamissonis than chemotype. For such relationships to be clarified, trials similar in design to those by Spitaler (2006) could also be conducted for this species, with the establishment of A. chamissonis from one source in a number of different Scottish environments. This would not be of economic benefit to the
Orkney region however and so such trials were not taken further in this study.
310
It should be nothed that care must be taken, when attempting to establish chemotypes based upon the variation of an individual or family of compounds.
Such variations may instead reflect differences that exist in the genome of the plant (Lawrence, 2000), or flexibility in active compound production in response to a variable environment. Further work is required to fully establish the biosynthetic relationships in both A. montana and A. chamissonis . Such work requires analysis of single populations, at identical stages of development, conducted alongside very careful recordings of the physical characteristics of each plant (in particular the leaves) so as to both clearly distinguish between sub species and to identify any relationship between structure and active compound production.
311 6. Discussion and conclusion
6.1 Aims and objectives
The research addressed the question ‘how does varied provence, environmental and agronomic factors, influence the quantity and range of active compounds found in A. montana and A. chamissonis’? The following discussion will critically examine the findings of the study in relation to these objectives and identify areas for future investigation.
6.2 Trial parameters
A range of potential crops are available to the farmer, but with each new option come the risk of failed crops, low yields, poor markets and lost revenue. This study aimed to generate information regarding crop suitability and potential yield so as to better inform farmers. Agricultural research as a whole aims to improve the economic effectiveness of agriculture either by improving the environments in which plants are grown (E) or by improving the genotypes (G). This study investigated aspects of both G and E in an attempt to determine which were most influential on yield. Although in theory it should be possible to distinguish G, E and
GE components of crop yield, in practice it can be impossible (Simmonds & Smart,
1979). This study was no different, but it did identify a number of key factors and possible avenues for further investigation.
Although this study was essentially a feasibility study of the growth of Arnica in
Orkney, the intention was that the knowledge gained would contribute to a better general understanding of the agronomic potential of both A. montana and A.
312 chamissonis as well as to how other Asteraceae crops might best be trialled. This would in turn be of interest to farmers from other areas who may be considering diversifying into this type of crop.
Arnica extract is an established herbal treatment which has meant that there is a well established demand for the dried flowers of Arnica (Burnie, unpublished). This is supported by a combination of the increased sales of herbal medicines (Lange
1998 and Kathe et al. 2002), the overexploitation of Arnica in the wild (Lange,
1998) and the unstable yields of cultivated plants (Cassells et al. 1999). The combination of market demand and decreased availability, led to Arnica being selected by the Agronomy Institute at Orkney College, as a trial crop, in 2002.
Trials of many crops have a natural focus on the weight or size of the ‘saleable’ part of the plant (e.g. the seed). However, because the study plant was being investigated for its potential as a high value extract crop, a quantitative analysis of the extract produced was required, alongside the more traditional yield measurements. Some research on the effect of agronomic variations on yield has already been conducted (e.g. Douglas et al. 2004) and there have also been investigations into the effect of variations in the natural habitat (Spitaler et al. 2006 and Willuhn et al. 1994). However, the research conducted for this thesis adopted a more novel approach and conducted both chemical and physical analyses of the effect of environmental and agronomic factors on yield. For example, although trials have shown that the application of fertiliser can increase flower yield, none have analysed the effect of fertiliser on the sesquiterpene lactone profile.
313 Complete chemical profiles and tests of purported medicinal properties have normally been conducted before growth trials of alternative crops are undertaken, and this was also the case for Arnica. The extract is an established anti- inflammatory agent (e.g. Wagner et al. 2004a, Wagner & Merfort, 2007, Siedle et al. 2004, Klaas et al. 2002) and so there was no requirement to repeat these experiments as part of this research. Sesquiterpene lactones have been identified as the anti-inflammatory agents within the extract (Wagner et al. 2004a), with thymol and essential oil fatty acids assisting penetration into the dermis (El Katten et al. 2001 and Wagner et al. 2004a). Quantitative and qualitative analyses of the levels of these compounds in the oil and the solvent extract were conducted in order to determine the quality of the flower material and the likely efficacy of its extract compared to commercially available flower material.
Extraction methods for both A. montana and A. chamissonis are well established as is the characterisation of the compounds within such extracts (e.g. Leven &
Willuhn, 1987, DAB10, 1991, Douglas et al. 2004) hence extensive extraction trials and confirmations of compound identification were not required. The focus of this research was rather on the effects of agronomic and environmental conditions on flower yield and associated overall extract quality, knowledge of which could then potentially support further enhancement of yields.
The above, combined with comparisons of the Orkney produced A. chamissonis and A. montana dried flowers with those available on the herbal market, and an analysis of the logistics involved in creating a marketable product, will here be used to critically assess what the ideal regime would be for Orkney grown Arnica
314 and whether it has any potential as a commercial high value extract crop in this region.
6.3 Oil yield, profile and sesquiterpene lactone content of A. chamissonis and A. montana
6.3.1 A. chamissonis
Levels of essential oil produced from the dried flowers of A. chamissonis were lower than that from A. montana flowers, stems and rhizomes. The oil from the former was also thicker and prone to emulsifying during the hydrodistillation process which was likely due to the higher n-alkane content (45%). A.chamissonis essential oil had higher levels of pinene, but fewer thymol compounds which are known penentration enhancers (El Katten et al. 2001). Twenty six compounds were identified in total (section 2.3.3) but the levels of oil produced were minimal.
As the oil is not used commercially this was not investigated further.
A. chamissonis had an average sesquiterpene lactone content of 0.17% w/w, which was considerably less than the 0.87% w/w reported by Willuhn et al. (1994).
A. chamissonis extract contained far fewer compounds than were found in other studies, for example nineteen were found by Level and Willuhn (1987), whilst the
Orkney produced flower extract contained just three: chamissonolide, arnifolin and
4-O-acetyl-6-desoxychamissonolide with the highest proportions found in the partially open flowers (Section 2.3.7.2). Such low levels of sesquiterpene lactones would imply that the flowers of A. chamissonis would not produce high quality extract.
315 As described in earlier chapters, there are Spanish and Central European chemotypes of A. montana, as well as five chemotypes of A. chamissonis.
Although the exact source of A. chamissonis used in this study could not be confirmed, it did not comply with any of the 5 chemotypes. The most similar was probably that which contained mostly chamissonolides and arnifolins (Type E)
(Willuhn et al. 1994). Attempts were made to grow A. chamissonis of a different chemotype, but no effect of seed source was found on the levels or types of sesquiterpene lactones in the extract of this species (section 5.3.3.5).
6.3.2 A. montana
Levels of essential oil produced were low for all A. montana material (section
2.3.2) but were highest for the rhizomes. The profile of the oils in the partially open flowers and the over flowers were quite different which suggests that the time of flower harvest will have a considerable impact on the quality of the extract.
There were 30 compounds found in the oil from the partially open flowers, compared to just 20 from the over flowers (section 4.3.3). The oil produced from partially open flowers was not as thick as that from the A. chamissonis flowers, with an n-alkane content of 24%, but the oil from the over flowers was much thicker, with a average n-alkane content of 53% .
The roots were highest in thymol based compounds (e.g. thymol methyl ether) which El Katten et al. (2001) identified as permeation enhancers of the sesquiterpene lactones. It would be unsustainable to harvest the roots every year but relatively high amounts were also found in the stem essential oil of A. montana . If such compounds from the stem material were included in the extract,
316 the permeation of the anti-inflammatory compounds might be enhanced. However, a balance would need to be struck between increasing thymol content and diluting the concentration of sesquiterpene lactones.
Attempts were made to boost the oil content by the addition of fertiliser, but this was not found to have any significant effect, only decreasing levels of heptacosane and Peak D in the partially open flowers. No effect of fertiliser was found on the oil content of over flowers and so this, combined with low levels over all, meant that oil content was not investigated further.
The total sesquiterpene lactone content in A. montana extract has been reported at levels ranging from 0.31% to 1.01% w/w (Willuhn et al. 1994). For Arnica extract to be effective as an anti-inflammatory, it has been recommended that the total sesquiterpene lactone content should contain not less than 0.4% w/w of total sesquiterpene lactones (European Pharmacopeia, 2000 and Willuhn & Leven,
1995). The average sesquiterpene lactone content of the Orkney grown material was 6.3 mg/ml (0.63% w/w) which was well within the required parameters. A comparison of French material (Organic Herb Trading Company, 2007) was made and the sesquiterpene lactone content of that was found to be 6.1 mg/ml (0.61% w/w) and although both were above the 0.4% w/w recommended for topical Arnica preparations, if the material was left until the flowers were ‘over’, then this content increased to 0.7% w/w. This indicates that leaving flowers later than recommended before harvesting improves the quality of the resulting extract and agrees with the results of Douglas et al. (2004).
317 Fourteen sesquiterpene lactones were found in the Orkney grown A. montana which was four more than those found by Douglas et al. (2004) in New Zealand, but was comparable to that found by Willuhn & Leven (1991). The most significant forms identified in all the aforementioned cases were the helenalin esters which have been described as the most active (Wagner et al. 2004b). The buds of A. montana contained significantly lower amounts of helenalin esters and higher amounts of the dihydrohelenalins (Section 2.3.5.2). As the flowers matured the levels of helenalin esters increased whilst the dihydrohelenalin esters decreased, suggesting that the dihydrohelenalins are transformed, perhaps via the actions of a dehydrogenase enzyme, during the flower’s development.
Siedle et al. (2004) and Klaas et al. (2002) demonstrated that the activity of the sesquiterpene lactones was dependent upon the ester group of the sesquiterpene lactones, with the unsaturated forms (i.e the methacryl and tigloyl esters) showing greater anti-inflammatory activity than the saturated forms. The extract from the
Orkney material contained higher levels of both of these compounds than the extract from the commercially available material. The greatest amounts in the
Orkney extract were the methacryl, isobutyryl, acetyl, and isovaleroyl esters respectively. In the commercial extract the same total numbers of sesquiterpene lactones were found, but for this material the greatest amounts were found of the isobutyryl, isovaleroyl, methacryl and acetyl esters, respectively. However, the comparison was just with one source of commercial flower material and more analyses would need to be completed for a more complete comparison to be conducted. Unfortunately it was not possible to source other commercial material which would have allowed this part of the study to be taken further. Regardless,
318 the results obtained show that the Orkney grown material is at least comparable to that available commercially.
Seed from CN seeds was found to produce plants that produced higher amounts of sesquiterpene lactones in their flowers, whilst that of Jelitto produced a wider range of sesquiterpene lactones (section 5.3.3.4). If a wider range of seed sources were screened it may be that a strain producing a higher yield of sesquiterpene lactones might be found. However, if such a strain were discovered, it too would require cultivation on a larger scale in order to produce sufficient quantities of seed. To this end, the agronomic regime was investigated further.
6.4 The effect of agronomic and environmental conditions on crop yield and quality
6.4.1 Environmental conditions
Although there were some concerns at the start of this study that wind might be detrimental to this crop by causing stem damage or lodging (Section 3.1.1), initial results indicated that this would not be the case, with the screened yield plots not displaying significantly higher yields than the unscreened yield plots. In subsequent years it was thought that this may instead reflect an ineffectiveness in the screens at blocking the Orkney winds, as a negative correlation was found for wind chill, average summer wind speed and to a lesser extent average wind speed
(Table 3.11), with both dry weight and number of flowers produced by A. chamissonis in both screened and non screened plots.
319 Whilst this might suggest that wind damage led to broken stems or damage to the plant, no such trend was detected for maximum wind speed or the number of gale days which would be expected to display the same trend. Negative correlations were found however, between the sesquiterpene lactone content of A. chamissonis and maximum wind and number of gale days. In contrast, a positive correlation was found between sesquiterpene lactone production in A. montana and the number of gale days, the opposite relationship to that found for A. chamissonis. This could be an indicator of a differing response to wind stress in the two species, which might be expected given the biogeography of the two species.
It may be that constant wind stress in A. chamissonis cause resources such as phosphoenol pyruvic acid, a common precursor to both the lignins and sesquiterpene lactones (Mooney, 1972) to be diverted away from sesquiterpene lactone production in order to strengthen the roots or stems. Only a few types of fungi can metabolize lignin and so this compound can also serve as a form of defence (Pew, 1967).
Lignans have been found in Arnica - in particular - the roots of A. chamissonis contain significant amounts of phenyl propanoids (Schmidt et al. 2006). The activation of the phenyl propanoid pathway in response to stress is widespread in plants and well known. For example, research by Cipollini (1997) showed that wind induced mechanical stimulation of bean plants, led to increased lignin levels, increased resistance to a fungal pathogen, increased resistance to bending and increased resistance to attack by spider mites. It may be that wind has stimulated
320 the production of lignins in A. chamissonis , which in turn may have caused a decrease in sesquiterpene lactone production.
However, it should be noted that a negative correlation was also found between the average summer wind speed and total sun hours and so it may be that correlations are artefacts of associations between environmental factors (section
3.3.5). In other words, it is hard to know whether a correlation of decreased sesquiterpene lactone production with summer wind is due to that environmental factor, or another factor such as sun or daylight hours. No relationship between wind and yield was found for A. montana.
Winter wind chill was associated with increased flower and dry weight production in both species, although it is likely this is due to wind chill being correlated with a yet undetermined environmental factor (section 3.3.5) because no Arnica plant material is above ground during these months, and no relationship was found with minimum winter temperature. These results could for example, reflect winter wind chill exerting an influence on surrounding weeds.
A positive correlation was found between total hours of sun and maximum number of sun hours with the dry weight and total number of flowers of A. chamissonis
(Section 3.3.5). When just the summer sun hours were taken into account, the relationship was not as strong which may indicate that the hours of sun that occur after flowering are important for rhizome reserves. If this is the case, then the process of ‘cutting back’ the leaf material after flowering, if it is to be employed,
321 should be left as long as possible so that the leaves can produce the reserves that are then stored in the rhizome for the following spring.
There was no significant relationship between the maximum hours of sun and the levels of sesquiterpene lactones in A. chamissonis (section 3.3.6) which could have indicated resources that would normally be used for defence have been allocated to increased biomass (eg. leaves).
Maximum rainfall had more of a negative effect on the number of flowers produced by A. montana than A. chamissonis (section 3.3.5) , although there were some indications (though not significant) that unscreened A. chamissonis was also affected by this factor. The strongest effect of total rainfall was on the number of flowers produced by unscreened A. chamissonis plants. The soil in Orkney is heavy and the soil profiles of the trial sites (section 3.3.1) suggest that waterlogging is common. It is possible that high maximum rainfall in Orkney is associated with waterlogging and that it is this that reduced the yield of A. montana.
No relationship between total hours of sun and flower number or flower dry weight was found for A. montana , although there were a number of negative correlations between the levels of individual helenalin esters present in A. montana extract and the number of summer sun hours. This relationship was particularly strong for isobutyryl helenalin, isovaleroyl helenalin and 2-methylbutyryrl helenalin. A similar relationship existed for maximum summer wind, though evidence of this was not evident if maximum wind over the year was used.
322
Overall, it seems that yield components such as flower dry weight and the number of flowers produced by A. chamissonis were more influenced by environmental factors than A. montana . In turn, the individual sesquiterpene lactone content of the extract of A. montana flowers appeared to be more influenced by environmental factors than that of A. chamissonis . In A. chamissonis the production of sesquiterpene lactones was fairly constant, whilst in A. montana the yield of these compounds varied considerably over the season. It may be that a factor other than fluctuations in environmental conditions, is limiting the production of flowers in A. montana . However, when these flowers are produced, the sesquiterpene lactone content of the flowers seemed to be heavily influenced by the environment.
It should be noted, that although some correlations were found between environmental factors and yield, a causative link does not necessarily follow
(Pakeman et al, 2008). It may be that a general decline in yield due to the age of the plants coincided with a decline in another environmental factor such as total sun hours. If this were the case, a link may appear to be significant, when it was in fact just a co-incidence. Hence these trial results would need to be compared with similar trials conducted in other environments for any causative relationship to be stated with any confidence. In addition, the weather recordings used in this study are monthly averages and provide a rather blunt analytical tool. For any of the above relationships to be described with confidence, a more intense monitoring programme would be required. Such monitoring would be beyond the original aims and objectives of this study. Instead, the results obtained here provide an insight
323 into what may have affected the yields of both species in Orkney, and serve as a starting point for further trials to identify the most appropriate agronomic regimes for Arnica . In particular, treatments that could help limit the effects of waterlogging and poor drainage.
6.4.2 Agronomic regimes
In addition to the environment, biological factors can also influence yield, in particular weeds, pests and pathogens (Leroux et al. 1996 and Abawi & Widmer,
2000). Normally all of these are unwanted by the farmer as they can reduce harvest index and/or contaminate the crop. Weeds require removal or spraying if they are not to contaminate the crop and reduce the harvest index. Their removal can lead to the root damage of the surrounding plants, but if they are left they can potentially decrease the amount of light available to the crop plant and serve as a reservoir of crop pests, particularly when their growing season is longer than that of the plant (Section 4.1). Disease also has the potential to reduce yield and alter the quality of the final product.
As the use of herbicides would be unlikely to be accepted by the predominantly organic market for Arnica , studies of the effects of manual weeding were conducted. A. chamissonis did not require weeding as it appeared to be highly competitive in the Orkney environment (section 3.3.4). Instead, this part of the study focused on the effect of weeds on A. montana which was not believed to compete effectively (Pegtel, 1994) and appeared to be at risk of being outcompeted by plants of similar growth habit (personal observation, section
3.3.4). The results of these trials demonstrated that whilst weeding appeared to
324 have no effect on the number of flowers produced or the dry weight of these flowers, this treatment seemed to stimulate the production of isobutyryl helenalin and isovaleroyl helenalin as well as total helenalins and total sesquiterpene lactone production. Unfortunately, in 2007 the weeding treatment was also strongly correlated with plant death (Section 4.3.1) which more than cancelled out any benefit of increased sesquiterpene lactone production. Although the quality of the floral extract is important, the sustainability of the flower harvest is vital. If weeding causes increased plant deaths as these results would seem to indicate, then it is important that an alternative form of weed control is applied.
As described earlier, perhaps due to its growth habit or maybe due to the production of alleleopathic compounds A. chamissonis did not require weeding.
Whilst allelopathy has been shown for a number of sesquiterpene lactones
(Picman 1986), it has not yet been demonstrated for Arnica extract. Regardless, whilst outcompeting weeds could potentially make A. chamissonis a more viable crop than A. montana , it was so effective at outcompeting weeds that it could have the potential to spread quickly beyond its area of cultivation. An invasive species is commonly defined as one that aggressively invades a continent to which it is not native (Middleton et al, 2008) and whilst this is unlikely whilst trials are conducted in the Orkney Islands, if further trials of this species are conducted, its invasive potential would have to be carefully monitored.
Flower dry weight production was weakly correlated with increased fertiliser
(section 4.3.2.2) but a balance has to be struck between stimulating the number of flowers produced and reducing the active compound production. It has been found
325 by Mihaliak & Lincoln (1989), for example, that production of secondary compounds was stimulated by limiting the application of fertiliser. If more flowers are produced, but the quality of such flowers is reduced, it may be that the additional cost of fertiliser does not compensate for a reduction in quality and associated value of the flowers.
Although the sample size was very small, fertiliser appeared to increase the levels of heptacosane and Peak D in the partially open flower oil and there was more perhydrofarnesylactone present in oils from non-weeded plants. No effect of fertiliser or weeding on flower weight or flower number was found for the over flowers, although an increase in caryophyllene production was associated with fertiliser treatment. Plots of A. montana that were fertilised also had higher amounts of isobutyryl helenalin and isovaleroyl helenalin as well as total amounts of helenalins and sesquiterpene lactones. Although Smallfield and Douglas (2008) found that fertiliser encouraged crown rot, no correlation between plant mortality and fertiliser treatment was found in this study.
An interaction effect between weeding and fertiliser was found for both isobutyryl and isovaleroyl helenalin. These compounds were found to the greatest extent in plots that were both fertilised and weeded, but in the absence of weeding most were found in plots that were not fertilised. This could have been a side effect of the shading by surrounding weeds (personal observation) which could in turn have limited the plant’s ability to produce secondary compounds. However, a negative relationship was found between hours of sun and the production of these compounds (section 3.3.6) which would suggest this was not the case.
326
Hand weeding would of course be expensive in terms of labour, and as shown, could lead to crop damage. Overall, the effect of weeding (or not weeding) was not hugely detrimental to the yield of A. montana , with flower dry weight, total numbers of flowers produced and levels of sesquiterpene lactones still present at acceptable levels when weeding did not take place. Taking into account the additional costs that would be involved in this treatment, as well as the increased number of plant deaths that seem to be linked with weeding, further trials of Arnica employing alternative methods of weed control were conducted.
Mulches were identified as a potentially viable alternative and hence were investigated further in this study. Initial trials employing polythene, sawdust and quarry fines as mulches, revealed that there was a significant effect of mulching on plant mortality (section 5.3.1.1), with the lowest number of plant deaths occurring in the plots with sawdust mulch. Although there exist no published results of field trials of mulches and Arnica with which to compare these results, the cause of decreased plant deaths was in this case believed to be due to the relative ease of removal of the weeds on this mulch compared to those under the other treatments.
This, combined with the results of the weeding trails (section 4.3.1), suggest that damage to the roots of A. montana is creating an opportunity for crown rot infection, with the consequential plant death. In order to confirm whether this was the case, this mulch was taken forward for further trials which also involved A. chamissonis . Although the latter did not require weeding, it was felt that as mulch can alter the soil it is covering (Himelick & Watson, 1990 and Hild & Morgan,
1993), the potential stress caused by such an alteration might stimulate the
327 production of additional sesquiterpene lactones within this species. These trials were combined with the use of ridges in an attempt to improve drainage and reduce the spread of crown rot.
In these subsequent trials, a highly significant effect of sawdust mulch on the average dry weight of A. montana flowers was found (F 1,13 =17.57, p=0.001,
Section 5.3.2.2), with those plants grown under mulch producing heavier flowers than those in the control plots. However, overall these plots did not produce more flowers and did not produce a greater overall total flower dry weight. This was unexpected, as normally an increase in flower weight with no corresponding decrease in flower number would mean an increased total flower dry weight.
However, when dry weight and total number of flowers were calculated on a per plant basis, there was no detectable effect of either treatment. This suggests that the indication of slightly higher mortality under mulch reported in section 5.3.1.2 may be interfering with the results.
For A. chamissonis a significant effect of mulching treatment on total flower dry weight was also found, but in this case a higher total dry flower weight was produced from the plants grown under mulch than that without (section 5.3.2.2).
This may be due to changes in the pH of the soil caused by the mulch (Himelick &
Watson, 1990). A. chamissonis has been reported to prefer alkaline soil (Plantlife,
2004) and it could be that increased acidification has created a stress which stimulated investment in flower production. It could of course be due to another factor such as a potentially increased reflection of photosynthetically active radiation (PAR) (section 5.4) as lighter mulches have been found to reflect more of
328 this than dark mulches (Decoteau et al. 1988). If this were the case, it would be expected that the increase in yield (section 5.3.2.2), and the lack of an effect on the level of sesquiterpene lactone production (section 5.3.3.3) found for the mulching treatment, would also be found when a higher number of either total, or summer sun hours occurred (section 3.3.6). For A. chamissonis , as for the mulching treatment, both total hours of sun and maximum sun hours were correlated with an increase in flower dry weight and number of flowers, whilst increased summer and total sun hours had no detectable effect on sesquiterpene lactone production either individually or in total (section 3.3.6). Only maximum hours of summer sun seemed to be linked to an increase in sesquiterpene lactone production in this species, and only for 4-O-acetyl-6-desoxychamissonolide and arnifolin. Hence it may be that an increase in reflection of PAR is responsible for the increase in total dry weight for this species, rather than any change in soil conditions, but this would need to be investigated further in future field trials, or in climate controlled conditions.
For A. montana, a small number of individual sesquiterpenes appeared to be affected by mulching treatment (section 5.3.3.2). For example, acetyl helenalin appeared to be significantly affected by both mulching and bed treatment, with most found in the extract from non mulched and ridged plots. There was however, no significant effect of either treatment on overall levels of either helenalin esters, dihydrohelenalin esters, or total sesquiterpene lactones, (section 5.3.3.2).
However, when the increased mass of flowers is taken into account, this does at least mean that no dilution of sesquiterpene lactone has occurred with the increased mass. If treatment were linked to an increase in reflection of PAR, then
329 it would be expected, as for A. chamissonis that this should be similar to the correlation with increased number of total and summer sun hours. However, in this case, no significant effect of total or summer sun hours on yield was evident
(section 3.3.5) nor was one apparent for most of the sesquiterpene lactones which apparently increased under mulch (section 3.3.6). Only methacryl helenalin was increased by both mulch and increased summer sun hours and so it is unlikely that reflection of PAR is significantly influencing the yield of A. montana .
The type of bed that A. montana was grown on had a significant effect on plant survival, with more plants dying on flat beds than on ridges (section 5.3.1.2).
Ridges require specific machinery in order for them to be created, which may limit the number of farmers that would be able to grow it in this way. However, the high number of plant deaths on flat beds suggests that it would be worth the extra cost, particularly where waterlogging could be a problem. As described above, this treatment had no significant effect on the quality of extract from either of the Arnica species.
6.5 The ‘ideal’ growing regime for Arnica grown in Orkney
6.5.1 A. chamissonis
On the basis of the research carried out for this study, the ideal regime for a high yield of A. chamissonis flowers would be on a site protected from the wind, but one in which winter frosts are still common. This area would not need to be weeded, though it would benefit from a sawdust mulch treatment. Flowers should be harvested at the partially open stage where possible. The growth habit of A. chamissonis allows it to easily out-compete local weeds and could potentially
330 return relatively high yields of flowers over many years. However, this also means it could potentially become a locally invasive species (Middleton et al. 2008). Any larger trials of this species would have to be carefully monitored to judge the scale of this threat.
The factors believed to be influencing production in both species of Arnica are presented in Tables 6.1 and 6.2. However, it should be noted that in order for A. chamissonis to become a successful alternative crop in Orkney, an alternative use for the extract would need to be identified, as it does not pass the regulations set down for by the Pharmacopeia (DAB 10, 1991) for minimum sesquiterpene lactone content.
Chamissonolide has been shown to have cytoprotective effects (Gertsch et al.
2003) but so far no commercial market for this compound has been established.
Due to the allergenic and/or cytotoxic properties of many sesquiterpene lactones, the formulation of new, bioavailable sesquiterpene lactones has been recommended (Ghantous et al. 2010). However, the purity of cytotoxic preparations needs to be much higher than that of an anti-inflammatory tincture.
If a cytotoxin of medical interest is found in A.chamissonis, when the genes encoding enzymes for such compounds have been identified they could potentially be transferred for expression in a plant such as tobacco, which produces a high amount of trichome exudate (Wagner, 1991). If such a procedure is successful, and the compounds of interest are successfully produced in this plant, it has been suggested they could then be removed from the leaves via solvent stripping,
331 without killing the plant (Chen et al. 2000). This process would be more efficient and more easily controlled than harvesting A.chamissonis and so it is unlikely that this species will be used as a source of medical extracts in future.
The lack of a commercially viable extract, combined with the high likelihood of A. chamissonis becoming a locally invasive species, and the difficulties that harvesting the smaller flowers present, lead to the conclusion that this species is not a suitable potential crop for Orkney.
6.5.2 A. montana
As described in section 3.1.1, a balance may need to be struck between the weight of flowers produced, and the quality of extract. This was thought to have been particularly likely if the ideal regime for the production of sesquiterpene lactones involved a stress that was detrimental to plant growth. Tables 6.1, 6.2 and 6.3 summarise the findings of the investigations into the effects of agronomic and environmental factors on yield.
These results taken together imply that the ideal regime for A. montana in Orkney would be for plants grown from CN Seed sourced material to be grown on ridges, mulched with sawdust and not fertilised. It should be on a relatively exposed site, without shelter, and preferably one that suffers from frost in winter. The flowers should be left as long as possible before harvesting.
Crown rot has not been reported in wild populations of A. montana , but this is common for many cultivated species. Wild plants rarely suffer from outbreaks of
332 epidemic diseases, for which their dispersed, discontinuous character is generally believed to be responsible. Agriculture reduces the genetic base, creates a large more or less continuously distributed population and as a result some pathogens do increase in number (Simmonds & Smart, 1979). Control of such pathogens from the outset is hence vital for yields to be maintained. If signs of crown rot appear in the cultivated A.montana , a new site should be found some distance away as this disease could potentially build up in the soil and spread to new sites.
Such sites should not be used again as Smallfield & Douglas (2008) found that this species will not regrow on infected sites.
It is likely that A. montana plots would need to be planted on a three year cycle as yields seem to consistently decline. It is possible that sales of the dried root would help compensate for the lack of a crop of flowers in the first year of the newly planted A. montana.
Table 6.1 – Factors correlated with flower dry weight and/or flower number in Arnica
Increases No effect Decreases Factor AM AC AM AC AM AC Minimum wind chill Average summer wind speed Average wind speed Total hours of sun Maximum hours of sun Summer sun hours Maximum daily rainfall Total rainfall Fertiliser Mulch Weeding Seed source
AM = A. montana AC = A. chamissonis. Fertiliser was not applied to the A. chamissonis trials
333
Table 6.2 – Factors correlated with sesquiterpene lactone production in Arnica
Increases No effect Decreases Factor AM AC AM AC AM AC Maximum wind Maximum summer wind Number of gale days Maximum hours of sun Summer sun hours Age of flowers Fertiliser Seed source Weeding Mulch
AM = A. montana AC = A. chamissonis. Weeding was not applied to the A. chamissonis trials
Table 6.3 – Factors correlated with mortalities in A. montana
Plant mortalities in A. montana Increase No effect or decrease Weeding Fertiliser Plastic mulch Ridges Mineral mulch Sawdust mulch Flat beds Seed source
It should be noted that wind chill and low temperatures were associated with increased yield in this study. On the basis of the predictions for future climate change, Orkney’s winter temperatures are likely to increase (Alcamo et al. 2007), it is recommended that future trials in Scotland should take place in environments of a higher elevation in the north and/or west of Scotland where there is a higher likelihood of colder winters. Although montane winters may also be subject to reduction in severity, in models of both low and high warming scenarios, low altitude grass/meadow regions (in which Arnica is found in central Europe) were both predicted to gain suitable climate space (Trivedi et al. 2008).
334
Another factor which should be taken into account when designing the ideal agronomic regime for Arnica is the presence of arbuscular mycorrhizal fungi
(AMF). The fungal net of hyphae associated with AMF increases the root surface and increases the plant’s access to water and minerals (particularly phosphorous).
Access to such nets has been known to increase resistance to pathogenic fungi
(Azcón-Aguila & Barea, 1997).
Recently published work by Jurkiewicz et al. (2010) has illustrated the importance of AMF for A. montana. They found that A. montana plants with no AMF produced less phenolic acids in roots and often less sesquiterpene lactones, had poorer photosynthetic performance and were generally outcompeted whist Ryszka et al.
(2010) found AMF in all surveyed wild populations of A. montana. AMF were not screened as part of this study, nor were measurements made of phenolic acids or photosynthetic performance. However, the levels of sesquiterpene lactones in
A.montana were comparable to that available commercially, which could indicate that AMF were present. However, in trials of this species with Dactylis glomerata
L. A.montana plants that were not inoculated were less competitive, but only at low fertiliser levels. This would correspond with the findings of this analysis where the lowest yields came from unfertilised plants. It is hence possible that AMF were not present in the Orkney trial sites.
AMF can colonize any suitable plant species (Mosse et al. 1981), but different plant species vary in their level of susceptibility (Gianinazzi-Pearson, 1984).
Although not possible for this study, it is recommended that any future trials should
335 also consider monitoring of AMF alongside agronomic regimes in both A. montana and A. chamissonis in order to determine whether AMF are present in both species and whether it is this that influences a plant’s susceptibility to crown rot.
6.6 The role of sesquiterpene lactones in Arnica
6.6.1 A. chamissonis
In this study, no effect of seed source was found for the levels of sesquiterpene lactones produced by A. chamissonis . Although it could be that each of the companies that provided A. chamissonis seed for this study had originally sourced it from the same supplier, it was felt this was unlikely due to the wide variety of suppliers sourced. This result could be employed alongside the high levels of natural variation found by Willuhn et al. (1994) to support the argument that the production of sesquiterpene lactones is strongly influenced by an as yet unidentified environmental or agronomic factor.
It has been found for some of the Asteraceae that the plants with the highest level of secondary metabolites were those whose growth was most limited by a lack of soil nutrients (Almeida-Cortezet al. 2004). This effect was even stronger where irradiance was also limiting. It may be that the factor that controls the sesquiterpene lactone content of A. chamissonis will also be one that limits its growth and this may limit its potential as an invasive species.
Further studies would be required in order to identify exactly what causes the variation in sesquiterpene lactone content among the natural populations of A. chamissonis (Willuhn et al. 2004) . However, it can be predicted with relative
336 confidence, that this species is very unlikely to be categorised as a potential alternative crop in Orkney in the near future.
However, the Highlands and Islands of Scotland present a diverse range of habitats and it is the recommendation of this study that further, small scale trials are conducted using a smaller number of sample plants in a wider range of habitats. The dried flower extract should then be analysed to determine whether either an increase in range and/or number of sesquiterpene lactones are produced in these environments. If they are, and the key stimulating factor is identified, then perhaps larger scale trials would be worth the investment. If no variation is detected, then further trials of this plant can be safely abandoned unless an alternative use for the extract can be found which, as described earlier (section
6.5.1), is unlikely.
6.6.2 A. montana
Although trials of the ARBO variety of A. montana showed that sesquiterpene lactone content was unaffected by altitude (Bomme & Daniel, 1994), Perry et al.(2009) found that extract of A. montana flowers from different altitudes had different levels of sesquiterpene lactones. It was hypothesised that the production of sesquiterpene lactones in A. montana might be under relatively rigid genetic control and so seed from a number of sources was trialled as per A. chamissonis.
A significant effect of seed source on amounts of sesquiterpene lactones produced was detected for A. montana (section 5.3.3.4) with the highest amounts found in the flowers of plants grown from seeds sourced from CN Seeds and with
337 chamissonolide, a sesquiterpene lactone more commonly produced by A. chamissonis (Willuhn & Kresken, 1981 and Leven & Willuhn, 1987), found in the extract of the flowers produced by plants from Jelitto sourced seed. The natural variation found in the plants sourced from different suppliers indicates that control of sesquiterpene lactone production may have been subject to natural selection and hence is heritable. This would in turn indicate that such control mechanisms are of importance to the plant and may be providing a benefit such as more efficient use of limited resources. Endemism has been linked to areas of poor soils but mild climates (e.g. Rodríguez-Sánchez et al. 2008), which indicates that strong selection for adaptations that allow plants to adapt to otherwise satisfactory environments is possible.
In studies of another Asteraceae ( Echinacea ), Wu et al. (2009) found strong correlation between the results of a chemotaxonomic study and a classification based on physical characteristics, whilst Flagel et al. (2008) in a study of neutral gene variation found close links between genetic structure and past rounds of
North American glaciations. These did not correspond to the chemotaxonomic studies by Wu et al. (2009) which suggests that chemotaxonomic endemism in the
Asteraceae is both possible, and capable of rapid establishment. This may also have occurred in the North American populations of A. chamissonis although further genetic screening (ideally involving markers of broad genetic coverage as it would appear that the phylogenetic signal within the Asteraceae may be weak on neutral genes), of this species would be required in order to confirm this hypothesis.
338 As emphasised earlier, correlation does not guarantee causation. It could be that the apparent higher sensitivity to environmental factors found for A. montana compared to A. chamissonis is actually due to another factor which is itself strongly influenced by environment. This factor would have to have a strong influence on yield and the processes within the plant controlling sesquiterpene lactone production, and yet not always be detectable. In addition, the results obtained in this study indicate that A. chamissonis would need to be relatively insensitive to this factor, whilst A. montana would be more sensitive. The only factors that could meet all of these criteria would be a disease such as crown rot, insect attack or the herbivory of small animals such as slugs or snails (larger animals would have been monitored more easily). The potential for one or all of these factors to influence yield and control of sesquiterpene lactone production in relation to the Orkney trial requires some consideration.
Sesquiterpene lactones are often deterrents for insect attack (Guillet et al. 2000) but the main insects found to attack Arnica (Tephritis arnicae and Melanagromyza arnicarum (both associated with increasing altitude in Europe, Scaltriti, 1985)) are not found in Orkney.
Scheidel & Bruelheide (1999) found that whilst the leaves of A. montana were highly palatable to slugs, predamaged leaves were less palatable than undamaged leaves. It may be that a form of chemical defence is induced by such damage, which may also occur in the flowers (the petals are often eaten though the rest of the flower is left intact). Sesquiterpene lactone production has been found to decrease with altitude as do the number of slugs and snails (Bruelheide &
339 Scheidel, 1999). As secondary compounds are expensive to the plant to produce, an inducible defence model would be an advantage. However, it is possible that it is not sesquiterpene lactones that are being produced to repel the slugs. Phenolics could be linked to defence in A. montana as they are generally avoided by slugs
(MØlgaard, 1986). In Arnica these compounds were found to increase with altitude whilst slug herbivory decreased (Spitaler et al. 2006). However, Bruelheide and
Scheidel (1999) found that although damage to Arnica by native slugs decreases with altitude, this may be due to effects of high altitudes on the slug life cycle, as when slugs were transported to high altitude populations of A. montana , the leaves were found to be highly palatable. The high levels of phenolics and the increase in ratio of quercetin to kaemferol, was initially purported by Spitaler et al. in 2006, to have a UV-B protective and radical scavenging function. However, more recent trials in growth chambers have attributed the changes in phenolic content to be stimulated by the lower temperatures at high altitude (Albert et al. 2009).
The correlation of decreased levels of sesquiterpene lactones with increased hours of sun, could reflect a potential decrease in herbivory and hence a decreased requirement for induced sesquiterpene lactone production. As shown in
Table 6.2, the other factors that led to an increase in sesquiterpene lactone production were those that were also linked to an increase in age (over flowers), repair (gale days) or increased nutrient availability (fertiliser and weeding). Those that negatively impacted on sesquiterpene lactone content were summer sun and summer wind, both of which would normally be linked to photosynthesis. Whilst these factors could also influence the local slug population it is unlikely that sesquiterpene lactones in A. montana are being induced to defend against slugs,
340 as their level would have been expected to increase in wet conditions and this was not the case.
The other possible stimulant of sesquiterpene lactone production was disease.
Crown rot was endemic throughout the A. montana plots and is believed to have been caused by the oomycetes Phytophthora and Pythium (DEFRA, 2005). One factor which would theoretically increase the levels of the oomycete infection found in the diseased plants would be waterlogging, whilst one that would decrease it would be summer drought (Watson, 1966). The environmental and agronomic factors that could exacerbate the former would be high rainfall and flat beds (both correlated with plant mortalities) whilst those that would influence the latter include summer sun and summer wind. Weeding was hypothesised to be damaging the stems and roots of the A. montana plants and creating sites for crown rot infection.
Smallfield & Douglas (2008) found that fertiliser encouraged crown rot, and whilst a correlation was not apparent in this study, it may be that the increase in sesquiterpene lactone production that was found was a response to infection.
Though if this were the case, it would be expected that a correlation between weeding and sesquiterpene lactone content would have also been found and this was not the case.
It should be stressed that the above hypotheses require further testing, as it is unwise to make statements describing ecological significance of compounds without extensive experimental data to support such claims (Simmonds et al).
However, the results of this research do provide some indications to the direction such future research should take.
341
6.7 The potential of Orkney grown A. montana as a commercial high value extract crop.
The agriculture sector is an important source of income in rural areas, even when it generates only a small proportion of this income directly. The retention of farmers in such areas can help keep small communities viable (de Janvry et al.
2002) and so any crop that could increase the sustainability of farming in such areas is worth investigation. However, rural areas are vulnerable to the effects of failed harvests and so it is important that trials of potential crops are conducted before diversification into a new species can be recommended.
In 2003 the total yield of A. montana was equivalent to 1407 kg/ha which was much higher than the 1000 kg/ha reported from German and Finnish trials
(Bomme and Daniel, 1994 and Galambosi et al. 1998 respectively). However, yield dropped each year to approximately 105 kg/h. As suggested earlier, it may be that this plant would work best as part of a rotation, with new plants established every few years. If this were to be implemented, the Orkney trials indicate that this should occur every three years (after which the yields dropped most sharply). After two years of A.montana harvests a yield of roughly 1100 kg/ha is possible. If the last harvest was also a total harvest (including rhizome material) this could increase the return for this crop and perhaps help compensate for the lack of flowering in year 1 of the next rotation of this crop.
The flowering of both A. montana and A. chamissonis is staggered (section 3.1.3) which in agricultural terms, requires either repeated harvesting which comes with
342 increased time and labour, or one or two harvesting ‘events’ which result in material of mixed maturity and hence reduced quality. Trials of non selective harvesting involving a cutter bar indicated that less frequent harvesting lead to greater harvest mass and a higher proportion of more mature flower material which was found in this study (section 2.3.5.2) and by Douglas et al. (2004) to be of higher quality. Douglas et al. (2004) also found that one simulated mechanical harvest of A. montana yielded twice as much plant material as two sessions of hand harvesting, although just over a third of this was stem material.
The stems of A. montana have been shown to contain low amounts of some sesquiterpene lactones (Douglas et al. 2004) although they are relatively high in thymol (a permeation enhancer, El Katten et al. 2001) (section 2.3.3). When these parts were removed by rubbing, the levels of the sesquiterpene lactones in these flowers and those that were hand harvested were comparable. This could mean that mechanical harvesting of A. montana would not result in a reduction in quality and hence would not require hand harvesting and the associated labour costs.
On the other hand, the equipment for mechanised harvesting is expensive, and because the flowers would all have to be dried at once, large drying facilities would be required to dry the flowers. The feasibility of A. montana as a crop hence is highly dependent upon the nature of the farmer. If it is a crofting operation, where
A. montana is grown on a steep slope in a remote area with limited facilities, then it may be that hand harvesting and staggered drying would be a more viable option. However, if the farmer was running a large scale operation, then
343 mechanical harvesting and forced air drying could potentially save on both labour resource and time.
Smallfield and Douglas (2008) found that 4-8 people could plant 800 to 1000 plants per hour, whilst 3 people and a 2 row mechanical punch planter could plant
3000 plants per hour (the additional people are required to correct those planted upside down). In trials of hand harvesting described in section 3.1.3 it was established that it would take approximately 3.5 hours to pick 1kg of A. montana and 9.2 hours to pick 1kg of A. chamissonis. 448 kg of A.montana would hence require 1568 hours of harvesting. At approximately £6 per hour (SAC, 2009), this would cost approximately £9,400 at current rates.
If the A. montana crop is re-established on a 3 yearly basis then the roots could also be sold to generate extra income. If the root has an average value of £60 per kg (Kathe, 2005) and if the dried root of each plant weighs approximately 10 g
(personal observation), then the root of each plant is worth £0.60. If 1 hectare was available to the farmer (10,000m 2) and 32 plants were planted per square metre, then one hectare of roots could potentially be worth £192,000. Land rental of one hectare was estimated by the SAC Farm Management Handbook (2009) to be approximately £620 per ha per year. The time taken to wash and dry the roots is hard to estimate, but as the process is likely to be similar to that of processing daffodil bulbs, then the SAC Farm Management handbook (2009) estimates 100h of labour for a hectare at about £9 per hour per person (£900). A modified potato harvester would be required (approximately £270 for one hectare). The roots
344 would then need to be dried, ideally using the same forced air drying facilities as the flowers.
Drying is best done locally and immediately after harvesting in order to ensure the quality of the dried material (Kathe, 2005). It also adds local value as dried flowers and roots are worth more than fresh and are easier to transport (ibid). If an investment in drying facilities was made, it should be noted that they could also be used for drying other material over the rest of the year. For example mushrooms could be dried when the facilites were not in use for drying A. montana.
If each kilogramme of dried flower material is worth approximately £20 per kg
(Burnie, unpublished), then one gram would be worth £0.02. If each plant produces an average of approximately 7 flowers, each weighing approximately
0.20g, then each plant would produce approximately 1.4 g or £0.028 worth of dried flower material. If 1 hectare was available to the farmer (10,000m 2) and 32 plants can be grown on 1m 2 (Burnie, unpublished), then £8,960 worth of dried flower material (448kg) could theoretically be produced each year.
The estimated costs and profit would depend upon where the farm was based, and other costs would include the application of sawdust, harvesting, drying, packaging and transport. The potential loss of profit due to disease is hard to predict, but this study shows they are likely to occur. If both planting out and harvesting of flowers and roots was done fully by hand, the cost of production would increase accordingly, as would the applications of treatments such as mulching. It should also be kept in mind that yields of Arnica can drop dramatically
345 and that the potential of Arnica as an alternative crop needs to be weighed against the value of other crops that could be grown on that land. It has been shown that
Arnica is best suited to land that is exposed and has sharp winter frosts. If this type of land is available, it is unlikely that it is currently in use for agriculture. In such a situation, A. montana has the potential to be a highly lucrative crop.
As described in chapter 1, there have been a number of phases of alternative agriculture since the early 1300s, but the common factor in each case was the breaking up of the larger farms facilitating increased diversification, vegetarianism and herbalism (Thirsk, 2000). However, many other factors influence this process, including the price of land, population size and the value placed on food production. Whilst it is possible that we are actually at the end of a phase of alternative agriculture (UK production of both wheat and Barley increased in 2008, whilst linseed production was low in both 2007 and 2008 (UK Agriculture, 2010)), the relationships between factors in this area are complex and hence it is very hard to state this with any confidence. The increased promotion of biofuels and biomass in response to depleting petrochemical stocks (EC, 2008) could equally indicate the start of a new phase of alternative agriculture.
Regardless, at the time of writing the UK is believed to be entering a recession which may have the potential to last some time (Lendman, 2010). During recessions the price of land normally decreases and the value of food increases
(Bernanke, 2000). It is likely that this, paired with a potential increased risk associated with the ‘untried’ will lead farmers away from alternative crops and back towards more conventional food production. Only those crops that have been
346 fully trialled and the extracts that have established medicinal properties have the potential to thrive in such economic climes. The results from this study indicate that A. montana has the potential to become one of those crops.
6.8 Recommendations
The results of this study have provided clear indicators as to where potential new avenues of research might be located. A. chamissonis produced a limited yield of sesquiterpene lactones both in quantitative and qualitative terms. It may be that by growing a small number of cloned plants in a wide variety of locations (carefully monitored for their tendency to spread), the factors that stimulate production of sesquiterpene lactones will be revealed.
This research showed that A. montana produced a good quality of extract in the
Orkney environment, but was vulnerable to crown rot and had a high number of plant deaths. Further research into the links between AMF, yield and competitiveness of this species are recommended to establish whether prior inoculation of the plants with AMF would be beneficial. In addition, trials of different colours of sawdust mulch, alongside careful monitoring of soil conditions, should be conducted in order to determine whether such a treatment would be beneficial to other high value flowering crops, and whether it was the change in soil conditions or the increase in blue light and PAR reflectance that increased yield in these trials.
347 6.9 Conclusions
The Asteraceae are one of the most genetically diverse and flexible families in the plant kingdom (Ingrouille & Eddie, 2006). Arnica is a circumboreal genus, found from the mountains of Europe, to Alaska, through Russia and as far south as New
Mexico. Yet there are very few Asteraceae which are grown as commercial crops.
This study has investigated the potential of A. montana and A. chamissonis in the
Orkney environment and has recommended that whilst A. chamissonis is not suitable, A. montana has potential as an alternative crop but high levels of disease present a potentially serious constraint. However, this study has also made recommendations for further research into the roles of sesquiterpene lactones in
Arnica which it is hoped can provide insights into ways in which their yields can be stabilised.
348
7. References
Abawi, G. S. and T. L. Widmer. 2000. "Impact of soil health management practices on soilborne pathogens, nematodes and root diseases of vegetable crops." Applied Soil Ecology 15(1):37-47.
Abraham, A., I. Kirson, E. Glotter and D. Lavie. 1968. "A chemotaxonomic study of Withania somnifera (L.) dun." Phytochemistry 7(6):957-962.
Adams, R. P. 2001. Identification of essential oil components by gas chromatography / quadrupole mass spectroscopy. IL: Allured.
Agriculture, UK. 2009. "Agristats." ed. Living Countryside.
Akhila, A, R.S. Thakur and S. P. Popli. 1987. "Biosynthesis of artemisinin in Artemisia annua ." Phytochemistry 26(7):1927-1930.
Albert, A, V Sareedenchai, W. Heller, H. K. Seidlitz and C. Zidorn. 2009. "Temperature is the key to altitudinal variation of phenolics in Arnica montana L. cv. ARBO." Oecologia (Berlin) 160(1):1-8.
Alcamo, J., J.M Moreno, B. Nováky, M. Bindi, R. Corobov, R.J.N Devoy, C. Giannakopoulos, E. Martin, J.E. Olesen and A. Shvidenko. 2007. "Europe: Climate Change 2007: Impacts, Adaptation and Vulnerability.Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change." eds. M.L. Parry, J.P. Canziani, P.J. van der Linden and C.E. Hanson. Cambridge, UK: Cambridge University Press.
Aldrich, R.J. 1984. Weed-Crop Ecology: Principles in Weed Management: North Scituate, Mass. (EUA). Breton.
Alfredo, P. P., C. A. Anaruma, A. C. S. Piao, S. M. A. Joao and R. A. Casarotto. 2009. "Effects of phonophoresis with Arnica montana onto acute inflammatory process in rat skeletal muscles: An experimental study." Ultrasonics 49(4-5):466- 471.
Almeida-Cortez, J. S., B. Shipley and J. T. Arnason. 2004. "Growth and chemical defense in relation to resource availability: tradeoffs or common responses to environmental stress?" Braz J Biol 64(2):187-194.
Alonso, D., M. C. Lazarus and L. Baumann. 2002. "Effects of topical arnica gel on post-laser treatment bruises." Dermatologic Surgery 28(8):686-688.
Alvarez-Castellanos, P. P. and M. J. Pascual-Villalobos. 2003. "Effect of fertilizer on yield and composition of flowerhead essential oil of Chrysanthemum coronarium (Asteraceae) cultivated in Spain." Industrial Crops and Products 17(2):77-81.
349 Anon. 2001. "Final report on the safety assessment of Arnica montana extract and Arnica montana ." International Journal of Toxicology 20(2):1-11.
APG, The Angiosperm Phylogeny Group. 1998. "An Ordinal Classification for the families of flowering plants." Annals of the Missouri Botanical Garden 85:531-553.
APG, The Angiosperm Phylogeny Group. 2002. "An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II." Botanical Journal of the Linnean Society 141:399-436.
Armbruster, D. A., M. D. Tillman and L. M. Hubbs. 1994. "Limit of detection (LOD) Limit of quantitation (LOQ) - Comparison of the empirical and the statistical, methods exemplified with GC-MS assays of abused drugs." Clinical Chemistry 40(7):1233-1238.
Azcón-Aguilar, C. and J. M. Barea. 1997. "Applying mycorrhiza biotechnology to horticulture: significance and potentials." Scientia Horticulturae 68(1-4):1-24.
Azizi, A, F. Y. and B. Honermeier. 2008. "Herbage yield, essential oil content and composition of three Oregano ( Origanum vulgare L.) populations as affected by soil moisture regimes and nitrogen supply." Industrial Crops and Products In Press, Corrected Proof.
Banthorpe, D.V. 1991. In Methods in Plant Biochemistry, eds. B.V. Charlwood and D.V. Banthorpe. London: Academic Press.
Beekman, A. C., H. J. Woerdenbag, W. van Uden, N. Pras, A. W. T. Konings, H. V. Wikstrom and T. J. Schmidt. 1997. "Structure - Cytotoxicity relationships of some helenanolide-type sesquiterpene lactones." Journal of Natural Products 60(3):252-257.
Belin, M, M Cambornac, J Crepin, L Flavet, F Guellec, G Redziniak and J R Robin. 1998. " Arnica chamissonis Less. an alternative to Arnica montana L. plant under threat, ." In 20th IFSCC Cong. Int. Fed. Soc. Cosmetic Chem. . London.
Bentham, G. 1873. "Compositae." In Genera Plantarum 2, ed. Bentham and Hooker. London.
Bernanke, B. 2000. Essays on the great depression. First Edition. Princetown: Princtown University Press.
Bezzi, A and G Ghidini. 1989. "Prime esperienze di coltivazione di Arnica montana L. sulle Alpi meridionali e sull'Appennino settentrionale." Annali ISAFA XI.
Bilia, A. R., M. C. Bergonzi, G. Mazzi and F. F.Vincieri. 2006. "Development and stability of semisolid preparations based on a supercritical CO 2 Arnica extract." Journal of Pharmaceutical and Biomedical Analysis 41:449-454.
Bloom, A. J. 1997. "Nitrogen as a limiting Factor: Crop Acquisition of Ammonium and Nitrate." In Ecology in Agriculture, ed. Louise E Jackson: Acadmic Press.
350 Bohlmann, J., G. Meyer-Gauen and R. Croteau. 1998. "Plant Terpenoid Synthases: Molecular Biology and Phylogenetic Analysis." Proceedings of the National Academy of Sciences of the United States of America 95(8):4126-4133.
Bolos y Vayreda, A. 1947. "El Arnica montana L. en la Penisula Iberica. Farmacognosia." Anales del Instituto Jose Celestino Mutis 7:145-151.
Bomme, U. and G. Daniel. 1994. "1st results on selection breeding of Arnica montana L." Gartenbauwissenschaft 59(2):67-71.
Bonamin, L. V. and P. C. Endler. 2010. "Animal models for studying homeopathy and high dilutions: Conceptual critical review." Homeopathy 99(1):37-50.
Borris, R. P. 1996. "Natural products research: perspectives from a major pharmaceutical company." Journal of Ethnopharmacology 51(1-3):29-38.
Braunschweig, Universitätsbibliothek. 2008. "ViFaPharm - Virtuelle Fachbibliothek Pharmazie." ed. M. Evers. Braunschweig: University Library Braunschweig.
Brinkhaus, B, J M Wilkens, R Lüdtke, J Hunger, C M Witt and S N Willich. 2006. "Homeopathic Arnica therapy in patients receiving knee surgery: Results of three randomised double blind trials." Complementary Therapies in Medicine 14:237- 246.
British Agrochemicals, Association. 1996. "Annual Review and Handbook."
British Herbal Medicine Association. 1996. The British Herbal Pharmacopoeia: The British Herbal Medicine Association.
Brown, M. M. 1903. "The Home Medicine Closet." The American Journal of Nursing 4(3):195-197.
Bruelheide, H. and U. Scheidel. 1999. "Slug herbivory as a limiting factor for the geographical range of Arnica montana ." Journal of Ecology 87:839-848.
Bryant, J. P., F. S. Chapin and D. R. Klein. 1983. "Carbon nutrient balance of boreal plants in relation to vertebrate herbivory." Oikos 40(3):357-368.
Bryant, J. P., K. F. Raffa and L. G. Barbara. 1995. "Chemical Antiherbivore Defense." In Plant Stems. San Diego: Academic Press.
Buckingham, J. 1994. Dictionary of Natural Products: v 1-7: CRC Press Inc.
Burnie, D. Unpublished Report, received 2004.
Buthuc-Keul, A.L and C Deliu. 2001. "Clonal propagation of Arnica montana L. a medicinal plant." In vitro Cellular and Developmental Biology - Plant 37(5):581- 585.
Caird, J .B. 1972. "Changes in the Highlands and Islands of Scotland 1951-1971."
351 Geoforum 12:5-36.
Camp, W. H. and C. L. Gilly. 1943. "Part II: Systematic Treatment, [ Arnica spathulata- Arnica sachalinensis ." Brittonia 4(3):457-493.
Cassells, A.C, C Walsh, M Belin, M Cambornac, J R Robin and C Lubrano. 1999. "Establishment of a plantation from micropropagated A. chamissonis a pharmaceutical substitute for the endangered A. montana ." Plant cell, tissue and Organ culture 56:139-144.
Celtest, Ltd. 2004. "Laboratory test report." ed. Orkney Direct Quarries. Bangor, Gwynedd.
Chakrabarti, J., S. J. Biswas and A. R. Khuda-Bukhsh. 2001. "Cytogenetical effects of sonication in mice and their modulations by actinomycin D and a homeopathic drug, Arnica 30." Indian Journal of Experimental Biology 39(12):1235-1242.
Chapin, F. S. III. 1991. "Integrated responses of plants to stress." Bioscience 41(1):29-36.
Chen, D.H., He-Chun Y. and Guo-Feng L. 2000. "Expression of a chimeric farnesyl diphosphate synthase gene in Artemisia annua L. transgenic plants via Agrobacterium tumefaciens -mediated transformation." Plant Science 155(2):179- 185.
Chavin, M.J.,P.S. Wakte and D.B. Shinde. 2010 “The anti-inflammatory activity of caryophyllene oxide from Annua squamosa L.” Phytomedicine 17: 149-151
Cipollini, D. F. 1997. "Wind-induced mechanical stimulation increases pest resistance in common bean." Oecologia 111(1):84-90.
Coley, P. D. 1988. "Effects of Plant Growth Rate and Leaf Lifetime on the Amount and Type of Anti-Herbivore Defense." Oecologia 74(4):531-536.
Connolly, J. D. and R. A. Hill. 1991. Dictionary of Terpenoids. London: Chapman and Hall.
Crofters Commission. 2007. "Annual Report." eds. D. Green and N. Reiter. Inverness: Crofters Commission.
Cronquist, A. 1955. "Compositae." In Vascular plants of the Pacific Northwest, eds. C.L Hitchcock, A Cronquist, M Ownbey and J Thompson, W. Princeton: Princeton University Press.
Cronquist, A. 1985. "History of generic concepts in the Compositae." Taxon 34(1):6-10.
DAB. 1964. Deutsches Arzneibuch. Stuttgart: Deutscher Apotheker Verlag.
352 DAB, 10. 1991. Deutsches Arzneibuch. 10th Edition. Stuttgart: Deutscher Apotheker Verlag.
Davies, D.B., D.J. Eagle and J.B. Finney. 2001. Resource Management: Soil. Revised Edition. Tonbridge: Farming Press.
De Janvry, A., E. Sadoulet and R. Murgai. 2002. "Chapter 3: Rural Development and Rural Policy." In Handbook of Agricultural Economics, eds. B. L. Gardner and G. C. Rausser.
Decoteau, D. R., M. J. Kasperbauer, D. D. Daniels and P. G. Hunt. 1988. "Plastic mulch color effects on reflected light and tomato plant growth." Scientia Horticulturae 34(3-4):169-175.
DEFRA. 2002. "Guidelines for farmers in NVZs - England." DEFRA.
DEFRA. 2005. "Arnica sample - final report." ed. Elizabeth Barron. York.
Delabays, N. and N. Mange. 1991. "La culture d' Arnica montana L. aspects agronomiqueset phytosanitaires." Rev Suisse Vitic Arboric Hortic 23:313-319. de la Fuente, E. B. de la, A. Gil, A. E. Lenardis, M. L. Pereira, S. A. Suárez, C. M. Ghersa and M. Y. Grass. 2003. "Response of winter crops differing in grain yield and essential oil production to some agronomic practices and environmental gradient in the Rolling Pampa, Argentina." Agriculture, Ecosystems & Environment 99(1-3):159-169.
Devon, T K and A I Scott. 1972. The Terpenes. London: Academic Press.
Douglas, James A, Bruce M Smallfield, Elaine J Burgess, Nigel B Perry, Rosemary E Anderson, Malcolm H Douglas and V. leAnne Glennie. 2004. "Sesquiterpene Lactones in Arnica montana : a Rapid Analytical Method and the Effects of Flower Maturity and Simulated Mechanical Harvesting on Quality and Yield." Planta Medica 70:166-170.
Dueck, Th A and J Elderson. 1992. "Influence of NH 3 and SO 2 on the growth and competitive ability of Arnica montana L. and Viola canina L." New Phytologist 122:507-514.
Duke, J.A, M.J. Bogenschutz-Godwin, J. duCellier and P.A. K Duke. 1985. Handbook of Medicinal Herbs: CRC Press.
Dunbabin, V. 2007. "Simulating the role of rooting traits in crop-weed competition." Field Crops Research 104(1-3):44-51.
EC, Council of the European Union -. 2008. "European Parliament legislative resolution of 17 December 2008 on the proposal for a directive of the European Parliament and of the Council on the promotion of the use of energy from renewable sources."
353 Eickmeyer, F. 2005. "Email regarding ARBO seeds ", ed. P Martin. Steinach.
Eissenstat, D. M. 1997. "Trade-offs in Root Form and Function." In Ecology in Agriculture, ed. L. E. Jackson: Academic Press.
El Katten, A F, C S Asbill, N Kim and B B Michniak. 2001. "The effects of terpene enhancers on the percutaneous permeation of drugs with different lipophillicities." International Journal of Pharmaceutics 215:229-240.
Ellenberger, A. 1998. "Assuming responsibility for a protected plant: Weleda’s endeavour to secure the firm’s supply of Arnica montana ." In First International Symposium on the Conservation of Medicinal Plants in Trade in Europe. Kew, UK: TRAFFIC Europe and The Royal Botanic Gardens, Kew, UK.
Ellis, S and A Mellor. 1995. Soils and Environment: Routledge.
Endt, D. Vom, J. W. Kijne and J. Memelink. 2002. "Transcription factors controlling plant secondary metabolism: what regulates the regulators?" Phytochemistry 61:107-114.
Essenberg, M. 2001. "Prospects for strengthening plant defences through phytoalexin engineering." Physiological and Molecular Plant Pathology 59:71-81.
Fabiszewski, J. and B. Wojtun. 2001. "Contemporary floristic changes in the Karkonosze Mts." Acta Societatis Botanicorum Poloniae 70(3):237-245.
FDA. 2005. "Code of Federal regulations." ed. FDA: US Government Printing Office
Fernandes, E. S., G. F. Passos, R. Medeiros, F. M. da Cunha, J. Ferreira, M. M. Campos, L. F. Pianowski and J. B. Calixto. 2007. "Anti-inflammatory effects of compounds alpha-humulene and (-)-trans-caryophyllene isolated from the essential oil of Cordia verbenacea ." European Journal of Pharmacology 569(3):228-236.
Fischer, N. H. 1978. "On the Biogenesis of Pseudoguaianolides." Rev. Latinoamer. Quim. 9:41-46.
Fischer, N. H. 1991. "Sesquiterpenoid Lactones." In Terpenoids, eds. B. V. Charlwood and D. V. Banthorpe: Academic Press.
Flagel, L. E., R. A. Rapp, C. E. Grover, M. P. Widrlechner, J. Hawkins, J. L. Grafenberg, I. Alvarez, G. Y. Chung and J. F. Wendel. 2008. "Phylogenetic, morphological, and chemotaxonomic incongruence in the North American endemic genus Echinacea." American Journal of Botany 95(6):756-765.
Flückiger, F. A. and D. Hanbury. 1874. Pharmacographia - A history of the principal drugs of vegetable origin met within Great Britain and British India. London: Macmillan and Co.
354 Forbes, J.C. and R.D. Watson. 1992. Plants in Agriculture. First Edition. Cambridge: University Press, Cambridge.
French, J. Unpublished personal communication (email) 2002
Fulton, R. M. J. 1979. "Soil Survey of Scotland." Director General of the Ordnance Survey, Southampton.
Galambosi, B., Z. Galambosi, K.P. Svoboda and G.G. Deans. 1998. "Blütenertrag und antioxidative Eigenshaften von Arnica montana L. kultiviert in Finnland." Drogenreport 11(19):10-13.
Gartlan, J.S., D.B. McKey, P.G. Waterman, C.N. Mbi and T.T. Struhsaker. 1980. A comparative study of the phytochemistry of two African rain forests. Biochem. Syst. Ecol. 8:401-422
GBIF. 2008. "Catalogue of Life: 2007 Annual Checklist: The Integrated Taxonomic Information System." Global Biodiversity Information Facility.
Geissman, T A. 1973. "The biogenesis of sesquiterpene lactones of the Compositae." In Terpeneoids: Structure, biogenesis and distribution, eds. V. C. Runeckles and T. J. Maybry. 6 Edition. New York and London: Academic Press.
Geran, R. I., N. H. Greenberg, M. M. MacDonald, A. M. Schumacher and B. J. Abbott. 1972. "Protocols for screening chemical agents and natural products against animal tumours and other biological systems." Cancer Chemother. Rep 3(3):1-103.
Gertsch, J., O. Sticher, T. Schmidt and J. Heilmann. 2003. "Influence of helenanolide-type sesquiterpene lactones on gene transcription profiles in Jurkat T cells and human peripheral blood cells: anti-inflammatory and cytotoxic effects." Biochemical Pharmacology 66(11):2141-2153.
Ghantous, A., H. Gali-Muhtasib, H. Vuorela, N. A. Saliba and N. Darwiche. 2010. "What made sesquiterpene lactones reach cancer clinical trials?" Drug Discovery Today In Press, Uncorrected Proof.
Gianinazzi-Pearson, V. 1984. "Host fungus specificity." In Genes involved in plant- microbe interactions, eds. L. Willmitzer, W. Schmalenbach and J. Schell. New York: Springer.
GLPMA. 1999. "The Good Laboratory Practice Regulations." ed. GLPMA: HMSO.
Goldberg, J. 1996. "Introduction." In The Black Death in England, eds. M. Ormrod and P.G Lindley: Stamford: P. Watkins.
Gottlieb, O. R. 1989. "Review article number 47: The role of oxygen in phytochemical evolution towards diversity." Phytochemistry 28(10):2545-2558.
Guillet, G, J. Harmantha, G. Waddell, T. B. J. R. Philogene and J T Arnason.
355 2000. "Synergistic insecticidal mode of action between sesquiterpene lactones and a phototoxin, alpha-terthienyl." Photochemistry and Photobiology 71(2):111- 115.
Hahnemann, S. 1993. Organon of Medicine: Cooper Pub.
Hall, J., B. Bealey and R. Wadsworth. 2006. "Assessing the risks of air pollution impacts to the condition of Areas/Sites of Special Scientific Interest in the UK." Joint Nature Conservation Committee.
Hamilton, E. 1852. " Arnica montana ." In Flora Homoeopathica. London: Leath and Ross.
Hanson, R. L., H. A. Lardy and S. M. Kupchan. 1970. "Inhibition of Phosphofructokinase by Quinone Methide and Alpha-Methylene Lactone Tumor Inhibitors." Science 168:378-382.
Harborne, J. B. 1988. Introduction to Ecological Biochemistry. 3rd Edition. London: Academic Press.
Harborne, J. B. 2000. "Arsenal for Survival: Secondary Plant Products." Taxon 49 (3, Golden Jubilee Part 1) : 435-449.
Harborne, J. B. and H. Baxter. 2001. The Chemical Dictionary of Economic Plants: John Wiley & Sons Ltd.
Hausen, B. M. 1996. "A 6-Year Experience With Compositae Mix." American Journal of Contact Dermatitis 7(2):94-99.
Herout, V. 1971. "Chemotaxonomy of the Family Compositae." In Pharmacognosy and Phytochemistry, eds. H. Wagner and L. Hörhammer. Munich: Springer-Verlag.
Herout, V and F Šorm. 1969. "Chemotaxonomy of the Sesquiterpenoids of the Compositae." In Perspectives in Phytochemistry - Proceedings of the Phytochemical Society Symposium, Cambridge, April 1968, eds. J. B. Harborne and T. Swain. London: Academic Press.
Herz, W. 1977a. "Biogenetic Aspects of Sesquiterpene Lactone Chemistry." Israel Journal of Chemistry 16:32-44.
Herz, W. 1968. "Pseudoguaianolides in Compositae." In Recent Advances in Phytochemistry, ed. T. J. Maybry. Amsterdam: Appleton Century Crofts.
Herz, W. 1977b. "Sesquiterpene Lactones in the Compositae." In The Biology and Chemistry of the Compositae, eds. V. H. Heywood and J. B. Harborne. London, New York, San Fransico: Academic Press.
Heywood, V. H. 1971. "Systematic survey of Old World Umbelliferae." In The Biology and Chemistry of the Umbelliferae, ed. V. H. Heywood. Reading: Academic Press.
356 Heywood, V .H., J. B. Harborne and B. L. Turner. 1977. "An Overture to the Compositae." In The Biology and Chemistry of the Compositae, eds. V H Heywood and J. B. Harborne. London, New York, San Francisco: Academic Press.
Hild, A.L and D.L Morgan. 1993. "Mulch effects on crown growth of five southwestern shrub species." J Environ Hortic 11(1):41-13.
Himelick, E. B. and G.W. Watson. 1990. "Reduction of oak chlorosis with wood chip mulch treatments." J. Aboric 16(10):275-278.
Huang, P.R., Yuan-Ming Y. and T.C. V Wang. 2004. "Potent inhibition of human telomerase by helenalin." Cancer Letters xx:1-6.
Hubbell, S. P., J. J. Howard and D. F. Wiemer. 1984. "Chemical Leaf Repellency to an Attine Ant: Seasonal Distribution Among Potential Host Plant Species." Ecology 65(4):1067-1076.
Hubbell, S. P., D. F. Wiemer and A. Adejare. 1983. "An antifungal terpenoid defends a neotropical tree ( Hymenaea ) against attack by a fungus growing ant (Atta )." Oecologia 60:321-327.
Iles, J. K. and M. S. Dosmann. 1999. "Effect of organic and mineral mulches on soil properties and growth of fariview flame red maple trees." Journal of Arboriculture 25(3):163-166.
Ingrouille, M. J. and B. Eddie. 2006. Plants: Evolution and Diversity: Cambridge.
Jackson, L. E. and G. W. Koch. 1997. "The Ecophysiology of Crops and their Wild relatives " In Ecology in Agriculture, ed. L. E. Jackson: Academic Press.
Jenelten, U. and U. Feller. 1992. "Mineral-nutrition of Arnica montana L and Arnica chamissonis ssp foliosa Maguire - differences in the cation acquisition." Journal of Plant Nutrition 15(11):2351-2361.
Jurkiewicz, A., P. Ryszka, T. Anielska, P. Waligorski, D. Bialonska, K. Goralska, M. Tsimilli-Michael and K. Turnau. 2010. "Optimization of culture conditions of Arnica montana L.: effects of mycorrhizal fungi and competing plants." Mycorrhiza 20(5):293-306.
Kahmen, S. and P. Poschlod. 2000. "Population size, plant performance and genetic variation in the rare plant Arnica montana L. in the Rhön, Germany." Basic Applied Ecology 1:43-51.
Kathe, W. 2005. "Conservation of Eastern-European Medicinal Plants; Arnica montana in Romania." In The Frontis Workshop on Medicinal and Aromatic Plants, eds. R. J. Bogers, L. E. Craker and D. Lange. Wageningen, The Netherlands: Springer.
Kathe, W. , S. Honnef and A. Heym. 2003. "Medicinal and Aromatic Plants in
357 Albania, Bosnia-Herzegovina, Bulgaria, Croatia and Romania; A study of the collection of and trade in medicinal and aromatic plants (MAPs), relevant legislation and the potential of MAP use for financing nature conservation and protected areas." eds. Wolfgang Kathe and Susanne Honnef: German Federal Agency for Nature Conservation (BfN).
Kathe, W. A., A. Hempfling and W. Fischer. 2002. "Healing Power from Nature." ed. WWF/TRAFFIC Europe-Germany: TRAFFIC International.
Kating, H., W. Rinn and G. Willuhn. 1970. "Composition of Arnica Species 3: Fatty Acids in Volatile Oils from Blossoms of different Arnica Species." Planta Medica 18:130.
Kelsy, R.G and F. Shafizadeh. 1980. "Glandular trichomes and sesquiterpene lactones of Artemisia nova (Asteraceae)." Biochemical Systematics and Ecology 8:371-377.
Kerr, A., S. Shackley, R. Milne and S. Allen. 1999. "Climate Change: Scottish implications scoping study." ed. The Scottish Executive Central Research Unit: Scottish Executive.
Killham, K. 1994. Soil Ecology. Cambridge: Cambridge University Press.
Kim, S. H, S. M Oh and T. S. Kim. 2005. "Induction of human leukemia HL-60 cell differentiation via a PKC/ERK pathway by helenalin, a pseudoguainolide sesquiterpene lactone." European Journal of Pharmacology 511:89-97.
Klaas, C. A., G. Wagner, D. Laufer, S. Sosa, R. Della Loggia, U. Bomme, H. L. Pahl and I. Merfort. 2002. "Studies on the anti-inflammatory activity of phytopharmaceuticals prepared from Arnica flowers." Planta Medica 68(5):385- 391.
Klimánková, E., K. Holadová, J. Hajslová, T. Cajka, J. Poustka and M. Koudela. 2008. "Aroma profiles of five basil ( Ocimum basilicum L.) cultivars grown under conventional and organic conditions." Food Chemistry 107(1):464-472.
Knuesel, O., M. Weber and A. Suter. 2002. " Arnica montana gel in osteoarthritis of the knee: An open, multicenter clinical trial." Advances in Therapy 19(5):209-218.
Kobiv, Y.I. 1992. „Ecology and population-ontogenetic peculiarities of Arnica montana L. (Asteraceae) in the Ukrainian Carpathians.” Ukrayins’kyi Botanichnyi Zhurnal 49, 46-51.
Kommission, E. 1984. "Monografie der Kommission E: Arnicae flos (Arnikablüten)." Bundesanz 228.
Kordali, S., A. Cakir, H. Ozer, R. Cakmakci, M. Kesdek and E. Mete. 2008. "Antifungal, phytotoxic and insecticidal properties of essential oil isolated from Turkish Origanum acutidens and its three components, carvacrol, thymol and p- cymene." Bioresource Technology 99(18):8788-8795.
358 Körner, C. 1999. Alpine Plant Life. Functional Plant Ecology of High Mountain Ecosystems. Berlin: Springer.
Kos, O., M. T. Lindenmeyer, A. Tubaro, S. Sosa and I. Merfort. 2005. "New Sesquiterpene Lactones from Arnica Tincture Prepared from Fresh Flowerheads of Arnica montana ." Planta Medica 71:1044-1052.
Kucera, M., O. Horacek, J. Kalal, P. Kolar, P. Korbelar and Z. Polesna. 2003. "Synergetic analgesic effect of the combination of arnica and hydroxyethyl salicylate in ethanolic solution following cutaneous application by transcutaneous electrostimulation." Arzneimittelforschung 53(12):850-856.
Kühn, J. G. 1858. Die krankheiten der Kulturegewachse, ihre ursachen und ihre Verhutung. Berlin.
Kupchan, S. M., M. A. Eakin and A. M. Thomas. 1971. "Tumor Inhibitors .69. Structure-Cytotoxicity Relationships among Sesquiterpene Lactones." Journal of Medicinal chemistry 14(12):1147.
Kupchan, S. M., D. C. Fessler, M. A. Eakin and T. J. Giacobbe. 1970. "Reactions of Alpha Methylene Lactone Tumor Inhibitors with Model Biological Nucleophiles." Science 168(3929):376-378.
Landon, J.R. 1988. "Towards a standard field assessment of soil texture for mineral soils." Soil survey and land evaluation 8(3):161-165.
Lange, D. 1998. "Europe's medicinal and aromatic plants: their use, trade and conservation." Cambridge: TRAFFIC International.
Lass, C., M. Vocanson, S. Wagner, C.M. Schempp, J.F. Nicolas, I. Merfort and S.F Martin. 2008. "Anti-inflammatory and immune regulatory mechanisms prevent contact hypersensitivity to Arnica montana L. ." Exp Dermatol 17:849-857.
Lawrence, B. M. 2000. "Essential oils: from Agriculture to Chemistry." The International Journal of Aromatherapy 10(3/4):82-98.
Lawrence, B. T. 1974. "Infraspecific differentiation in the genus Mentha ." In 6th International Essential Oil Congress. San Francisco.
Lee, K.H., I.H. Hall, E.C. Mar, C. O. Starnes and S. A. Eigebaly. 1977a. "Sesquiterpene Antitumor Agents: Inhibitors of Cellular Metabolism." Science 196(4289):533-536.
Lee, K. H. , T. Ibuka, R. Yang Wu and T. A. Geissman. 1977b. "Antimicrobial Agents. 2. Structure-Antimicrobial Activity Relationships Among Sesquiterpene Lactones and Related Compounds." Phytochemistry 16:1177-1181.
Leistra, M and J.J.T.I Boesten. 1989. "Pesticide contamination of groundwater in Western Europe." Agriculture, Ecosystems & Environment 51:7-20.
359 Lendman, S. 2010. "The Recession is Over, the Economic Depression Just Beginning." The Market Oracle.
Leroux, G. D., D. L. Benoît and S. Banville. 1996. "Effect of crop rotations on weed control, Bidens cernua and Erigeron canadensis populations, and carrot yields in organic soils." Crop Protection 15(2):171-178.
Lessing, C. F. 1831. "Synanthereae Rich. In: A. von Chamisso & D. F. L. von Schlechtendal. De plantis in expeditione speculatoria Romanzoffiana, etc." Linnaea 6:83–170, 209–260, 501–528.
Leven, W. and G. Willuhn. 1987. "Sesquiterpene lactones from Arnica chamissonis Less. VI identification and quantitative determination by High- Performance Liquid and Gas Chromatography." Journal of Chromatography 410:329-342.
Levin, D. A. 1976. "The Chemical Defenses of Plants to Pathogens and Herbivores." Annual Review of Ecology and Systematics 7:121-159.
Liebman, M. and E. R. Gallandt. 1997. "Many little hammers: Ecological Managment of Crop-Weed Interactions." In Ecology in Agriculture, ed. Louise E Jackson: Academic Press.
Linnaei, C. 1753. Species Plantarum.
Lorenz, O.A. 1978. "Potential nitrate levels in edible plant parts " In Nitrogen in the environment, eds. D.R Nielsen and J.G. MacDonald. San Diego: Academic Press.
Luijten, S.H., A. Dierick, J.G.B. Oostermeijer, L.E.L. Raijmann and H.C.M. Den Nijs. 2000. "Population size, genetic variation and reproductive success in a rapidly declining, self incompatible perennial ( Arnica montana ) in The Netherlands." Conservation Biology 14(6):1776-1787.
Luijten, S.H., J. Gerard, M .J. B Oostermeijer, N. C. van Leeuwen and D. Nijs 1996. "Reproductive success and clonal genetic structure of the rare Arnica montana (Compositae) in the Netherlands." Plant systematics and Evolution 201(1-4):15-30.
Lussignoli, S, S. Bertani, H. Metelmann, P. Bellavite and A. Conforti. 1999. "Effect of Traumeel S, a homeopathic formulation, on blood-induced inflammation in rats." Complement Ther Med 7(4):225-230.
Lyß, G., A. Knorre, T. J. Schmidt, H. L . Pahl and I. Merfort. 1998. "The Anti- inflammatory sesquiterpene lactone helenalin inhibits the transcription factor NF- KB by directly targeting p65." The Journal of Biological Chemistry 273(50):33508- 33516.
Mabry, T.J. and F. Bohlman. 1977. “Summary of the Chemisty of the Compositae” In Chemisty and Biology of the Compositae (V.H. Heywood, J.B. Harbourne and B.L. Turner eds.) New York – London, Academic Press.
360 Macêdo, S. B., L. R. Ferreira, F. F. Perazzo and J. C. Tavares Carvalho. 2004. "Anti-inflammatory activity of Arnica montana 6cH: preclinical study in animals." Homeopathy 93:84-87.
Macias, F.A., J.C.G. Galindo, J.M.G. Molinillo and D. Castellano. 2000. “Dehydrozaluzanin C: a potent growth regulator with potential use as a natural herbicide template. Phytochemistry 54: 165-171.Maguire, B. 1943. "A monograph on the genus Arnica ." Brittonia 4:386-510.
Mahmoud, S. S and R. B Croteau. 2001. "Metabolic Engineering of Essential Oil Yield and Composition in Mint by Altering Expression of Deoxyxylulose Phosphate Reductoisomerase and Menthofuran SynthaseMetabolic Engineering of Essential Oil Yield and Composition in Mint by Altering Expression of Deoxyxylulose Phosphate Reductoisomerase and Menthofuran Synthase." Proceedings of the National Academy of Sciences of the United States of America 98(15):8915-8920.
Malarz, J., A. Sojakowska, B. Dohnal and W. Kisiel. 1993. "Helenalin acetate in in vitro Propagated plants of Arnica montana ." Planta Medica 59:51-53.
Martin, P. Unpublished. "Review of 2003 Trials and plantings." Agronomy Institute for Northern Temperate Crop Research, Orkney College UHI.
Martin, P. and J. French. Unpublished. Agronomy Institute Annual Report 2002.
McLafferty, F. W. 1959. "Mass Spectrometric Analysis. Molecular Rearrangements." Annal Chem 31(1):82-87.
Mead, R., R.N. Curnow and A.M. Hasted. 1993. Statistical Methods in Agriculture and Field Biology. Second Edition. Padstow, Cornwall: Chapman and Hall.
Merfort, I. 1988. "Acetylated and other flavonoid glycosides from Arnica chamissonis ." Phytochemistry 27(10):3281-3284.
Merfort, I. 1992. "Caffeoylquinic acids from flowers of Arnica montana and Arnica chamissonis ." Phytochemistry 31(6):2111-2113.
Merfort, I. 2002. "Review of the analytical techniques for sesquiterpenes and sesquiterpene lactones." Journal of Chromatography A 967(1):115-130.
Metoffice, UK. 2009. "Climate averages for Kirkwall, 1971-2000." UK Metoffice.
Middleton, B., A. Jorgensen, S. Erik and F. Brian. 2008. "Invasive Species." In Encyclopedia of Ecology. Oxford: Academic Press.
Mihaliak, C. A. and D. E. Lincoln. 1989. "Plant biomass partitioning and chemical defense - response to defoliation and nitrate limitation." Oecologia 80(1):122-126.
Mølgaard, P. 1986. "Food plant preferences by slugs and snails: A simple method to evaluate the relative palatability of the food plants." Biochemical Systematics and Ecology 14(1):113-121.
361 Money, T. and M. K. C. Wong. 1996. "A Formal, Enantiospecific Synthesis of Pseudoguaianolides." Tetrahedron 52(18):6307-6324.
Mooney, H. A. 1972. "The Carbon Balance of Plants." Annual Review of Ecology and Systematics 3:315-346.
Mosse, B., D. P. Stribley and F. Letacon. 1981. "Ecology of mycorrhizae and mycorrhizal fungi." Advances in Microbial Ecology 5:137-210.
Nes, W. R. and M. L. MacLean. 1977. Biochemistry of Steroids and other Isopentenoids: University Park Press.
Newman, D. J. and G. M. Cragg. 2006. "Natural Products from Marine Invertebrates and Microbes as Modulators of Antitumor Targets." Current Drug Targets 7(3):279-304.
Oberbaum, M., N. Galoyan, L. Lerner-Geva, S. R. Singer, S. Grisaru, D. Shashar and A. Samueloff. 2005. "The effect of the homeopathic remedies Arnica montana and Bellis perennis on mild postpartum bleeding - A randomized, double-blind, placebo-controlled study-Preliminary results." Complementary Therapies in Medicine 13(13):87-90.
Organic Herb Trading Co. 2006. "Email". Elizabeth Barron.
Pakeman, R. J., C.L. Reid, J. J. Lennon and M. Kent. 2008. "Possible interactions between environmental factors in determining species optima." Journal of Vegetation Science.
Palaiseul, J. 1976. Grandmother's Secrets. First Edition: Penguin Books Ltd.
Parolo, G., G. Rossi, and A. Ferrarini. 2008. “Towards improved species niche modelling: Arnica montana in the Alps as a case study. Journal of Applied Ecology. 45(5):1410-1418.
Pegtel, D. M. 1994. "Habitat characteristics and the effect of various nutrient solutions on growth and mineral nutrition of Arnica montana L. grown on natural soil." Vegetatio 114:109-121.
Perry, N. B., E. J. Burgess, M. A. R. Guitian, R. R. Franco, E. L. Mosquera, B. M. Smallfield, N. I. Joyce and R. P. Littlejohn. 2009. "Sesquiterpene Lactones in Arnica montana : Helenalin and Dihydrohelenalin Chemotypes in Spain." Planta Medica 75(6):660-666.
Petersen, M. 2007. "Current status of metabolic phytochemistry." Phytochemistry 68(22-24):2847-2860.
Pew, J C. 1967. "Lignin: an inviting storehouse." Agricultural Science Review 5(4):1-7.
Pharacopoeia, European. 2002. European Pharacopoeia. 4th Edition: EDQM.
362 Pharmacopoeia, European. 2000. European Pharmacopoeia. 3rd Edition: EDQM.
Picman, A. K. 1986. "Biological activities of sesquiterpene lactones." Biochemical Systematics and Ecology 14(3):255-281.
Plantlife. 2004. "Leafy Arnica - Arnica chamissonis ."
Poplawski, J., M. Holub, Z. Samek and V. Herout. 1971. "Arnicolides - Sesquiterpenic Lactones from the Leaves of Arnica montana L." Collection Czechoslov. Chem. Commun. 35:2189-2199.
Quirk, L. and T. Somers. 2009. "Grapevine Management Guide." National Wine and Grape Industry Centre.
Raison, J. v, J. Heilmann, I. Merfort, T. J. Schmidt, F E Brock, Walter Leven, U Bomme and R Bauer. 2000. "Arnika - Arzneipflanze mit Tradition und Zukunft." Zeitschrift für Phytotherapie 21:39-54.
Recio, M. C., R. M. Giner, L. Uriburu, S. Manez, M. Cerda, J. R. De la Fuente and J. L. Rios. 2000. "In vivo activity of pseudoguaianolide sesquiterpene lactones in acute and chronic inflammation." Life Sciences 66(26):2509-2518.
Reichardt P.B., F.S. Chapin, J.P. Bryant, B.R. Mattes and T.P. Clausen. 1991. “Carbon/nutrient balance as a predictor of plant defence in Alaskan balsam poplar: Potential importance of metabolite turnover. Oecologia, 888(3) 401-406.
Rickett, H. W. 1973. Wild flowers of the United States. New York: McGraw Hill Book Company.
Robinson, H. E. 1981. "A revision of the tribal and subtribal limits of the Heliantheae (Asteraceae)." Smithson. Contributions to Botany 51.
Robles, M., M. Aregullin, J. West and E. Rodriguez. 1995. "Recent Studies on the Zoopharmacognosy, Pharmacology and Neurotoxicology of Sesquiterpene Lactones." Planta Medica 61:199-203.
Rodríguez-Sánchez, F., R. Pérez-Barrales, F. Ojeda, P. Vargas and J. Arroyo. 2008. "The Strait of Gibraltar as a melting pot for plant biodiversity." Quaternary Science Reviews 27(23-24):2100-2117.
Roki, D., N. Menkovic, D. Savikin-Fodulovic, M. Ristic and D. Grubisic. 2001. "Flavonoids and essential oil in flower heads of Introduced Arnica chamissonis ." Journal of Herbs, Spices and Medicinal Plants 8(4):19-27.
Ruzicka, L. 1953. "The Isoprene Rule and the Biogenesis of Terpenic Compounds." Experientia 9(15):357-396.
Rybalko, K.S., E.A. Trutneva and P.N. Kubal'chich. 1965. "A crystalline substance isolated from Arnica foliosa ." Aptechnoe delo 14(4):32-33.
363 Ryszka P., J. Błaszkowski, A. Jurkiewicz and K. Turnau. 2010. Arbuscular mycorrhiza of Arnica montana under field conditions—conventional and molecular studies. 20 (8): 551-557
SAC. 2006. "Soil report." Scottish Agricultural College.
SAC. 2009. Farm Management Handbook: SAC.
Saker, M.M., S.M. Shanab, M. Dessouky and M. R Khater. 2000. " In vitro studies on Ambrosia maritima : II-Fungal elicitor enhances sesquiterpene lactones accumulation." Arabian Journal of Biotechnology 3(2):225-232.
Scaltriti, G. P.. 1985. "Gli insetti delle piante officinali. i nota l' Arnica montana L. e due suoi fitofagi specifici." Giornale di Zoologia 68:355-364.
Scheffer, F. and P. Schachtschabel. 1979. Lehrbuch der Bodenkunde. Stuttgart: Ferdinand Enke Verlag.
Scheidel, U. and H. Bruelheide. 1999. "Selective slug grazing on montane meadow plants." Journal of Ecology 87:828-838.
Scheidel, U., S. Rohl and H. Bruelheide. 2003. "Altitudinal gradients of generalist and specialist herbivory on three montane Asteraceae." Acta Oecologica- International Journal of Ecology 24(5-6):275-283.
Schmidt, B., D. M. Ribnicky, A. Poulev, S. Longendra, W. T. Cefalu and I. Raskin. 2008. "A natural history of botanical therapies." Metabolism Clinical and Experimental 57(suppl 1):3-9.
Schmidt, T. J. 1996. "Helenanolide type sesquiterpene lactones Part 1. Conformations and molecular dynamics of helenalin, its esters and 11,13-dihydro derivatives." Journal of Molecular Structure 385:99-112.
Schmidt, T. J. 1997. "Helenanolide-Type Sesquiterpene Lactones- III. Rates and Stereochemistry in the Reaction of Helenalin and Related Helenanolides with Sulfhydral Containing Biomolecules." Bioorganic and Medicinal Chemistry 5(4):645-653.
Schmidt, T. J., U. Bomme and A. W. Alfermann. 1998. "Sesquiterpene Lactone Content in Leaves of in vitro and Field Cultivated Arnica montana ." Planta Medica 64:268-270.
Schmidt, T. J., U. Matthiesen and G. Willuhn. 2000. "On the stability of sesquiterpene lactones in the officinal Arnica tincture of the German Pharmacopoeia." Planta Medica 66(7):678-681.
Schmidt, T. J., S. Stausberg, J. Von Raison, M. Berner and G. Willuhn. 2006. "Lignans from Arnica species." Natural Product Research 20(5):443-453.
Schroder, E. and I. Merfort. 1991. "Thermospray liquid-chromatographic mass-
364 spectrometric studies of flavonoid glycosides from Arnica montana and Arnica chamissonis extracts." Biological Mass Spectrometry 20(1):11-20.
Schwabe, A. 1990. "Syndynamic processes in Nardo callunetea communities: Changes in fallow land after renewed cattle grazing and life history of Arnica montana L. ." Carolinea 48:45-68.
Scottish Crofting Foundation. 2009. "Crofting connections."
Scottish Executive. 2003. "Action Introduced to reduce nitrate pollution." Scottish Exectutive.
Scottish Government. 2003. "Scottish Agricultural Census summary sheets by geographic area." The Scottish Government.
Scottish Office, Agriculture and Fisheries Dept. 1992. "Agriculture in Scotland." HMSO.
Seaman, F. C. and V. A. Funk. 1983. "Cladistic Analysis of Complex Natural Products: Developing Transformation Series from Sesquiterpene Lactone Data." Taxon 32(1):1-27.
Siedle, B., A.J. Garcia-Pineres, R. Murillo, J. Schulte-Monting, V. Castro, P. Ringeler, C.A. Klass, F.B. Da Costa, W. Kisiel. and I. Merfort. 2004. "Quantitative structure activity relationship of sesquiterpene lactones as inhibitors of the transcription factor NF-kappa B " Journal of Medicinal Chemistry 47:6042-6054.
Simmonds, M. S. J. 2003. "Flavonoid-insect interactions: recent advances in our knowledge." Phytochemistry 64(1):21-30.
Simmonds, M. S. J., C. M. Rai and M. Cecilia. 2006. "Chapter 13 The search for plant-derived compounds with antifeedant activity." In Advances in Phytomedicine: Elsevier.
Simmonds, N.W. and J. Smart. 1979. Principles of Crop Improvement. 2nd Edition: Blackwell.
Small, E. and P.M. Catling. 1999. Canadian Medicinal Crops: NRC Research Press.
Smallfield, B.M. and M.H. Douglas. 2008. " Arnica montana a grower's guide for commercial production in New Zealand." New Zealand Arnica Grower's Group.
Soil Association. 2009. "What's your poison? Chlormequat."
Spence, M. 1908. "The Climate of Orkney." Journal of Scottish Meteorological Society 14(25):238-255.
Spettoli, E., S. Silvani, P. Lucente, L. Guerra and C. Vincenzi. 1998. "Contact dermatitis caused by sesquiterpene lactones." Am J Contact Dermat 9(1):49-50.
365 Spitaler, R, P. D. Schlorhaufer, E. P. Ellmerer, I. Merfort, S. Bortenschlager, H. Stuppner and C. Zidorn. 2006. "Altitudinal variation of secondary metabolite profiles in flowering heads of Arnica montana cv. ARBO." Phytochemistry 67:409- 417.
Strykstra, R. J., D. M. Pegtel and A. Bergsma. 1998. "Dispersal distance and achene quality of the rare anemochorous species Arnica montana L.: implications for conservation." Acta Botanica Neerlandica 47(1):45-56.
Sy, L.K., N.Y. Zhu and G. D. Brown. 2001. "Syntheses of dihydroartemisinic acid and dihydro-epi-deoxyarteannuin B incorporating a stable isotope label at the 15- position for studies into the biosynthesis of artemisinin." Tetrahedron 57(40):8495- 8510.
Takeya, K., M. Itoigawa and H. Furukawa. 1983. "Positive Inotropic Effect Of Helenalin, A Sesquiterpene Lactone, On Guinea-Pig Myocardium." Chem. Pharm. Bull 31:1719-1725.
Taylor, R. S. L. and G. H. N. Towers. 1998. "Antibacterial constituents of the Nepalese medicinal herb, Centipeda minima." Phytochemistry 47(4):631-634.
Thirsk, J. 2000. Alternative Agriculture. First Edition. Oxford: Oxford Univesity Press.
Thompson, M. 2009. "Herbal Medicine " In She knows.
Thomson, C. and E. Antill. 1769. "An Essay on the Cultivation of the Vine, and the Making and Preserving of Wine, Suited to the Different Climates in North-America. By the Hon. Edward Antill, Esqr. Communicated by Mr. Charles Thomson." Transactions of the American Philosophical Society 1:117-197.
Tivy, J. 1990. Agricultural Ecology. London: Longman Scientific and Technical
Tosco, U. 1978. Mountain Flowers. London: Orbis Publishing.
Towrie, S. 2008. "Orkneyjar - the heritage of the Orkney Islands."
Trivedi, M. R., M. D. Morecroft, P. M. Berry and T. P. Dawson. 2008. "Potential effects of climate change on plant communities in three montane nature reserves in Scotland, UK." Biological Conservation 141(6):1665-1675.
Tsai, L., R. J. Highet and W. Herz. 1969. "The mass spectra of pseudoguaianolides related to helenalin." J Org Chem 34(4):945-948.
USDA, Natural Resources Conservation Service. 2007. "National Soil Survey Handbook." ed. Natural Resources Conservation Service.
USDA, Natural Resources Conservation Service. 2010. "Plant Profile - Arnica chamissonis." In PLANTS Profile, ed. Natural Resources Conservation Service USDA: USDA.
366 van den Berg, L. J. L. , C. J. H. Peters, M. R. Ashmore and J G M Roelofs. 2008. "Reduced nitrogen has a greater effect than oxidised nitrogen on dry heathland vegetation." Environmental Pollution xx:1-11.
Verdeguer, M., M. A. Blazquez and H. Boira. 2009. "Phytotoxic effects of Lantana camara , Eucalyptus camaldulensis and Eriocephalus africanus essential oils in weeds of Mediterranean summer crops." Biochemical Systematics and Ecology 37(4):362-369.
Verpoorte, R. and J. Memelink. 2002. "Engineering secondary metabolite production in plants." Current Opinion in Biotechnology 13:181-187.
Victor, F. C., D. E. Cohen and N. A. Soter. 2010. "A 20-year analysis of previous and emerging allergens that elicit photoallergic contact dermatitis." J Am Acad. Dermatol 62(4):605-610.
Vyvyan, J. R. 2002. "Allelochemicals as leads for new herbicides and agrochemicals." Tetrahedron 58(9):1631-1646.
Waddell, T. G., P. H. Gebert and D. L. Tait. 1983. "Michael-type reactions of tenulin, a biologically-active sesquiterpene lactone." Journal of Pharmaceutical Sciences 72(12):1474-1476.
Waggoner, P. E., P. M. Miller and H. C. De Roo. 1960. "Plastic mulching - principles and benefits." The Connecticut Agricultural Experiment Station Bulletin 634:1-44.
Wagner, G. J. 1991. “ Secreting glandular Trichomes: More than just hairs”. Plant Physiology. 96:675-679
Wagner, G. J., E. Wang and R. W. Shepherd. 2004. "New approaches for studying and exploiting an old protuberance, the plant trichome." Annals of Botany 93(1):3- 11.
Wagner, H. 1977. "Pharmaceutical and economic uses of the Compositae." In The Biology and Chemistry of the Compositae, eds. V. H. Heywood and J. B. Harborne. London, New York, San Francisco: Academic Press.
Wagner, S, F. Kratz and I. Merfort. 2004b. "In Vitro Behaviour of Sesquiterpene Lactones and Sesquiterpene Lactone Containing Plant Preparations in Human Blood, Plasma and Human Serum Albumin Solutions." Planta Medica 70(3):227- 233.
Wagner, S. and I. Merfort. 2007. "Skin penetration behaviour of sesquiterpene lactones from different Arnica preparations using a validated GC-MSD method." Journal of Pharmaceutical and Biomedical Analysis 43(1):32-38.
Wagner, S., A. Suter and I. Merfort. 2004a. "Skin penetration studies of Arnica preparations and of their sesquiterpene lactones." Planta Medica 70:897-903.
367 Watson, A.G. 1966. "Seasonal variation in inolculum potentials of spermosphere fungi." New Zealand Journal of Agricultural Resources 9:956-963.
Weston, L. A. 1996. "Utilization of allelopathy for weed management in agroecosystems." Agronomy Journal 88(6):860-866.
(WHO), World Health Organisation. 1990. "Public Health Impacts of Pesticides used in Agriculture." ed. WHO. Geneva.
Wichtl, M. 1987. "DAB9 - Kommentar." Stuttgart: Deutscher Apotheker Verlag.
Williams, D. G. 1996. The Chemistry of Essential Oils: An Introduction for Aromatherapists, Beauticians, Retailers and Students: Micelle Press.
Willuhn, G. 1998. "Arnica flowers; pharmacology, toxicology and analytics of the sesquiterpene lactones as their main active substances." In Phytomedicines of Europe. Chemistry and Biological Activity., eds. L D Lawson and R Bauer. Washington DC: American Chemical Society.
Willuhn, G. 1972 (a). "Untersuchungen über die inhaltsstoffe von Arnica -arten." Planta Medica 21:221-245.
Willuhn, G. 1972 (b). "Untersuchungen über die inhaltsstoffe von Arnica -arten." Planta Medica 21.
Willuhn, G. 1972 (c). "Untersuchungen über die inhaltsstoffe von Arnica -arten." Planta medica 22(1):1-33.
Willuhn, G., I. Junior and D. Wendisch. 1986. "Demethoxyencecalin And Thymol Derivatives From Arnica-Sachalinensis." Planta Medica 5:349-351.
Willuhn, G. and J. Kresken. 1981. "Sesquiterpene lactones from Arnica chamissonis ." Planta Medica 42(2):107-107.
Willuhn, G., J. Kresken and D. Wendisch. 1983. "Sesquiterpenlactone aus Arnica chamissonis .3. 4-0-acetyl-6-desoxychamissonolid und 6-0-propionyl-11,13- dihydrohelenalin." Planta Medica 47(3):157-160.
Willuhn, G. and W. Leven. 1991. "Zur qualitativen und quantitativen Analyse der Sesquiterpenlactone von Arnikablüten DAB 9." PZ-Wissenshaft 136:32-39.
Willuhn, G. and W. Leven. 1995. "Qualität von Arnikazubereitungen." Deutsche Apotheker Zeitung 135(21):33-36.
Willuhn, G., W. Leven and C. Luley. 1994. "Arnikablüten DAB 10 - Untersuchungen zur qualitativen und qualitativen variabilität des sesquiterpenlactongehaltes der offizinellen Arnikadrogen." Deutsche Apotheker Zeitung 134(42):49-59.
Willuhn, G., I. Merfort, C. M. Paßreiter and T. J. Schmidt. 1995. "Chemistry and
368 Systematics of the Genus Arnica ." In Compositae Systematics, eds. D J N Hind, C Jeffrey and G V Pope: Kew Publishing.
Woerdenbag, H.J., I. Merfort, T.J. Schmidt, C.M. Passreiter, G. Willuhn, W.V. Uden, N. Pras and A.W.T Konings. 1995. "Decreased helenalin induced cytotoxicity by flavonoids from Arnica as studied in a human lung carcinoma cell line." Phytomedicine 2:127-132.
Wu, L., P. M. Dixon, B. J. Nikolau, G. A. Kraus, M. P. Widrlechner and E. S. Wurtele. 2009. "Metabolic Profiling of Echinacea Genotypes and a Test of Alternative Taxonomic Treatments." Planta Medica 75(2):178-183.
Zakharov, P. I., R. I. Evstratova and K.S. Rybalko. 1971. "A mass-spectometric study of Arnifolin, a new sesquiterpene lactone from Arnica montana and A.foliosa ." Chemistry of Natural Compounds 7:567-569.
369