Birds exploit herbivore-induced plant volatiles to locate ...digital.csic.es/bitstream/10261/124406/4/EcolLet 16(11) 1348-1355 (… · 1 1 Birds exploit herbivore-induced plant volatiles
Post on 14-Jun-2020
0 Views
Preview:
Transcript
1
Birds exploit herbivore-induced plant volatiles to locate herbivorous 1
prey 2
3
Luisa Amo1*, Jeroen J. Jansen2, Nicole M. van Dam3, Marcel Dicke4 & Marcel E. 4
Visser1 5
6
1 Department of Animal Ecology, Netherlands Institute of Ecology (NIOO-KNAW) , 7
P.O. Box 50, 6700 AB Wageningen, The Netherlands 8
2 Department of Analytical Chemistry, Institute for Molecules and Materials, Radboud 9
University Nijmegen, 6525 AJ Nijmegen, The Netherlands 10
3 Department of Ecogenomics, Institute of Water and Wetland Research (IWWR), 11
Radboud University Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands 12
4 Laboratory of Entomology, Wageningen University, P.O. Box 8031, 6700 EH 13
Wageningen, The Netherlands 14
15
16
* Correspondence and present address: Luisa Amo. Departamento de Ecología 17
Funcional y Evolutiva, Estación Experimental de Zonas Áridas, CSIC. Carretera de 18
Sacramento, s/n E-04120, La Cañada de San Urbano, Almería, Spain. Telephone: 0034 19
950281045; Fax: 0034 950277100. E-mail: luisa.amo@eeza.csic.es 20
21
22
Running title: Birds smell herbivore-induced plant volatiles 23
Keywords: Multitrophic interactions, induced indirect plant defense, insect herbivores, 24
insectivorous birds, Parus major, foraging, apple trees, avian olfaction 25
2
Type of article: Letters 26
Number of words in the abstract: 149 27
Number of words in the main text: 4551 28
Number of references: 41 29
Number of figures: 5 30
Number of tables: 0 31
Number of text boxes: 0 32
33
Statement of authorship: L.A., M.E.V. & M.D. planned the experiments and wrote the 34
manuscript. L.A. carried out the experiments and took visual and chemical 35
measurements of trees. N.M.v.D. advised on volatile sampling and chemical compound 36
analysis. L.A. performed the statistical analysis of data, except the PLS-DA which was 37
carried out by J.J.J. 38
39
3
40
Arthropod herbivory induces plant volatiles that can be used by natural enemies of the 41
herbivores to find their prey. This has been studied mainly for arthropods that prey upon 42
or parasitize herbivorous arthropods but rarely for insectivorous birds, one of the main 43
groups of predators of herbivorous insects such as lepidopteran larvae. Here, we show 44
that great tits (Parus major) discriminate between caterpillar-infested and uninfested 45
trees. Birds were attracted to infested trees, even when they could not see the larvae or 46
their feeding damage. We furthermore show that infested and uninfested trees differ in 47
volatile emissions and visual characteristics. Finally, we show, for the first time, that 48
birds smell which tree is infested with their prey based on differences in volatile profiles 49
emitted by infested and uninfested trees. Volatiles emitted by plants in response to 50
herbivory by lepidopteran larvae thus not only attract predatory insects but also 51
vertebrate predators. 52
53
54
4
55
Among the most exciting interspecific relationships mediated by chemical cues 56
are multitrophic interactions involving plants, herbivorous arthropods and carnivorous 57
arthropods. When a plant is attacked by herbivorous arthropods, it induces a defense 58
response. The metabolites that the plant produces as a defense may directly affect the 59
performance of the herbivorous arthropod (induced direct defense) by, for example, 60
inhibiting feeding behavior of insects, decreasing digestibility or intoxicating the insect 61
(Schoonhoven et al. 2005). Furthermore, it has been proposed that the volatiles that 62
plants emit upon attack by herbivorous arthropods have an indirect defense function by 63
attracting carnivorous enemies of the herbivores (induced indirect defense, Dicke et al. 64
1990a; Turlings et al. 1990; Turlings & Tumlinson 1992; Vet & Dicke 1992). In doing 65
so, plants may reduce the damage by the herbivore, and thus, can enhance their fitness 66
(van Loon et al. 2000; Fritzsche Hoballah & Turlings 2001; Schuman et al. 2012). 67
The phenomenon of herbivore-induced emission of volatile organic compounds 68
by plants has mainly been studied considering insect enemies of the herbivores (see 69
Mumm & Dicke 2010; Dicke & Baldwin 2010 for reviews). However, many bird 70
species, such as the great tit, Parus major, are voracious predators of herbivorous 71
insects such as lepidopteran larvae, including the winter moth (Operopthera brumata, 72
Lepidoptera, Geometridae). Because the nestling period of the great tit coincides with 73
the peak occurrence of winter moth larvae, birds can greatly reduce the number of 74
lepidopteran larvae feeding on trees (Mols & Visser 2002). Predation of winter moths 75
by great tits has been found to decrease herbivore damage to trees (Mols & Visser 2002; 76
Van Bael et al. 2003; Mäntylä et al. 2011). This leads to increased growth and reduced 77
mortality of the trees (Marquis & Whelan 1994; Sipura 1999; Mäntylä et al. 2011). 78
Therefore, plants that are infested by herbivorous insects could benefit from the 79
5
attraction of insectivorous birds. Furthermore, insectivorous birds could also benefit 80
from the use of plant cues that enhance their chances to find their herbivorous prey. The 81
prey itself usually emits low amounts of cues thereby reducing detection by predators 82
(Rowland et al. 2008), whereas information emitted by the plant may be much easier to 83
detect due to the considerably larger biomass of plants compared to herbivores (Vet & 84
Dicke 1992). Previous evidence suggests that birds are attracted to trees infested by 85
lepidopteran larvae, without the need to see larvae or their damage on leaves (Mäntylä 86
et al. 2004, 2008a,b), but the mechanism underlying the attraction remains unknown. 87
Here, we present experiments aimed to elucidate whether birds are attracted to trees 88
infested by herbivorous prey and to explore the mechanism underlying such attraction 89
in the system: great tits - winter moths - apple trees. 90
In order to examine whether birds are attracted to trees infested by lepidopteran 91
larvae, we performed a two-choice experiment in an aviary (Fig. 1) containing two 92
types of apple trees, one control and one experimental tree. We investigated the first 93
visit and the proportion of visits by the birds to the tree that was experimentally infested 94
with winter moth larvae. We tested whether great tits preferred a) trees infested with 95
larvae, b) trees containing damaged leaves, from which larvae had been removed, or c) 96
trees infested by larvae, from which both larvae and the damaged part of each leaf had 97
been removed. If birds are able to use larva-induced tree volatiles we expected birds to 98
prefer the tree infested by lepidopteran larvae, even when larvae or the damaged leaves 99
had been removed before the choice experiment. 100
Next, we analyzed the mechanism responsible for the preference of great tits for 101
infested trees. We examined whether great tits were attracted to a) chemical cues, b) 102
visual cues, c) chemical & visual cues of apple trees infested with lepidopteran larvae, 103
from which damaged parts of leaves and the larvae themselves had been removed just 104
6
prior to the experiment. To further explore the potential cues used by the birds we 105
quantified the chemical and visual differences between infested and uninfested trees. 106
We expected that infested trees differed from the uninfested trees in the visual and 107
chemical cues that they emitted. Predatory arthropods are known to discriminate 108
between infested and uninfested trees based on the chemical cues that plants emit 109
(Schoonhoven et al. 2005). Despite the fact that birds are traditionally considered to 110
primarily use vision, recent evidence suggests that olfaction may be used more often 111
than previously thought, also in foraging contexts (e.g. Nevitt 2011). Therefore, we also 112
expect birds to, at least partly, rely on chemical cues to discriminate between infested 113
and uninfested trees. 114
115
MATERIALS AND METHODS 116
Insect species 117
In November 2006 and 2007 winter moth females (Operopthera brumata L.) were 118
captured in several deciduous forests to the west of Arnhem (05º48´E, 51º59´N), The 119
Netherlands. Females were kept individually in 50 mL falcon tubes (30 mm O.D., 115 120
mm length) to lay their eggs. Clutches were kept in petri dishes in outdoor conditions 121
until March, when eggs were transferred to climate cabinets (SANYO Incubator MIR-122
553) and maintained at 12 °C. Fresh young leaves of peach and apple trees were 123
provided to the containers to ensure that newly hatched larvae would have food. Larvae 124
were reared on these leaves until they reached the fifth larval instar (L5). 125
126
Tree species 127
From the beginning of April 2007 and 2008 we placed thirty-five 1.5 m tall apple trees, 128
Malus silvestris Miller (variety De Costa), planted in 40 L pots inside a greenhouse for 129
7
two weeks before the development of leaves. After leaf development, trees were moved 130
outdoors to habituate to experimental conditions. We separated control and 131
experimental trees several meters apart (minimum 10 meters) to avoid interactions 132
between them. 133
Three days before the experiment, we individually placed 30 winter moth (O. 134
brumata) larvae (L5) inside clip-cages (Ø = 250 mm) on each tree assigned the 135
“infestation” treatment. In this way, larvae could eat the leaf but could not move from 136
one leaf to another. We used 30 larvae because we wanted to mimic a natural situation, 137
where birds can find larvae in some but not in all tree leaves. Uninfested trees were 138
maintained without larvae. 139
140
Bird species 141
We used naïve captive adult great tits, Parus major, housed individually in 0.9 m × 0.4 142
m × 0.5 m cages. Birds were one year old and all of them were hand-reared since they 143
had been 10 days old; therefore, they did not have any previous experience in foraging 144
among trees. 145
Before the experiments, all birds were habituated to the aviary by releasing them 146
once during one hour inside the aviary without apple trees. In all experiments, we 147
removed the food from the cages that housed the experimental birds one hour before 148
each trial to ensure that the birds were motivated to search for larvae during the 149
experiment. After the trial, the bird was captured with a net and returned to its cage. 150
Birds did not show signals of stress during the trials and when they were returned to 151
their cages they immediately resumed their normal behavior. All experiments were 152
carried out under license of the Animal Experimental Committee of the KNAW (DEC 153
protocol no CTE 07.01). 154
8
We used thirty-eight adult great tits to test whether birds were attracted to trees 155
infested with lepidopteran larvae, and thirty-five other birds to examine the mechanism 156
underlying the discrimination between infested and uninfested trees. A repeated 157
measure design was used in both experiments. All birds were tested in the three 158
treatments in a randomized order. Only one trial was conducted per bird per day, and 159
there was at least one day without testing between trials. Before these experiments, 160
birds were trained five times to acclimatize them to the aviary and to allow them to find 161
larvae in the apple trees. During habituation trials, the mesh was partially removed (in 162
the experiment to assess the attraction to infested trees) or the door opened (in the 163
experiment to unravel the mechanism underlying the attraction to infested trees) to 164
allow birds to have access to both trees that were equally infested with larvae in these 165
trials. To maintain the birds’ interest to search for larvae during the trials, between each 166
trial of the experiment, we performed one habituation trial with each bird to allow it to 167
eat larvae from trees at both locations of the trees within the aviary simultaneously. 168
169
Experimental design 170
Experiments were performed in late April and early May in 2007 and 2008 in two 171
outdoor Y-shaped aviaries built with mesh screens (mesh size 1.3 cm) (Fig. 1). Each 172
branch of the aviary was 2.5 m x 2 m x 2 m (l x w x h). The central branch was closed 173
72 cm near the intersection with the other two branches. The aviary contained three 174
perches, one near each tree and the third in the middle of the aviary. 175
176
Experiment 1: attraction to infested trees 177
In this experiment, two apple trees were placed at the end of the branches of the aviary, 178
separated 4.40 meters from each other. One of the trees was uninfested and the other 179
9
tree infested. The infested tree had one of the following treatments: a) ‘Caterpillar’ 180
(infested tree with 30 lepidopteran larvae and damaged leaves), b) ‘Damaged leaves’ 181
(infested tree with damaged leaves from which the larvae had been removed), or c) 182
‘Previously infested’ (infested tree from which larvae and damaged parts of leaves had 183
been removed). The larvae and damaged parts of leaves were removed just before the 184
trials. We cut the damaged parts of leaves in the “previously infested” treatment. We cut 185
the part of the leaf that had been in contact with the larvae to remove not only the visual 186
damage but also any chemical compound left by larvae such as feces. We removed the 187
clip-cage containing the larvae by cutting the part of the leaf where the clip-cage was 188
located (about half a leaf). We also cut a similar part of the same number of leaves in 189
the uninfested trees and in the infested trees in the other treatments. Trees were covered 190
with protective mesh, to prevent birds from eating the larvae in the ‘caterpillar’ 191
treatment. We used 18 different pairs of trees. 192
193
Experiment 2: mechanism underlying the attraction to infested trees 194
Infested and uninfested trees (see above) were obtained as previously described. We 195
removed larvae and damaged leaves from infested trees and removed a similar number 196
of leaves in control, uninfested trees. Therefore, infested trees were similar to those of 197
the “previously infested” treatment in the former experiment. In this experiment, two 198
apple trees where placed at the end of each branch of the aviary and we thus had four 199
apple trees in the aviary (Fig. 1). Each pair of trees was located in a compartment with 200
two parts. One of the parts of the compartment contained a methacrylate door, and the 201
tree could be seen but not smelled. The other part of the compartment contained a cloth 202
(cotton) door, so the tree could be smelled but not seen by the birds. One of the pairs of 203
trees was control and the other one experimental. In the control pair of trees, trees were 204
10
always uninfested. The experimental pair of trees could have one of the following 205
treatments: a) ‘chemical’, b) ‘visual’, c) ‘chemical and visual’. In the ‘chemical’ 206
treatment, the tree that could be seen was uninfested and the tree that could be smelled 207
was infested. In the ‘visual’ treatment, the tree that could be seen was infested and the 208
tree that could be smelled was uninfested. In the ‘chemical and visual’ treatment, both 209
trees were infested, and therefore, birds could smell and see an infested tree. We used 210
12 different groups of 4 trees. 211
212
Experimental procedure 213
Trials were performed between 09:00 and 17:00 and under sunny and warm conditions 214
(mean + SE temperature = 20 + 1 °C) to avoid variation in the emission of volatiles due 215
to differences in ambient conditions such as temperature (Vallat et al. 2005). On each 216
test day, a new control and experimental tree or pair of trees was placed in each aviary. 217
We randomized the place of the trees (right or left) as well as the aviary (number one or 218
number two) among trials. Trees were tested with several birds (mean and median = 6 219
birds, from 2 to a maximum of 7 birds). We recorded the behavior of birds during 30 220
minutes using a video camera. An observer, blind to the treatments, analyzed the video 221
tapes and recorded the first tree inspected by the bird and the number of visits to each 222
tree during 30 minutes in the two experiments. We calculated the proportion of visits to 223
the control and experimental tree. 224
We analyzed the first choice as well as proportion of visits to the experimental 225
tree by using generalized linear mixed models fit by the Laplace approximation, with 226
these variables following a binomial distribution with logit function. The individual as 227
well as the tree pair were included in the model as random factors. Treatment, tree 228
location (left or right), aviary location (left or right) and order of trial were included in 229
11
the model as fixed factors, when relevant. Day, hour, hour2 and temperature were also 230
included in the initial models but were removed when they were not significant. 231
Treatment effect was calculated by comparing the models with and without the 232
treatment with ANOVA. Analyses were performed with the Statistical package R 2.15.1 233
(R Development Core Team 2012). Cases where birds did not visit any tree where 234
excluded from the analysis (two cases in the experiment to assess the attraction to 235
infested trees, and 23 cases in the experiment to disentangle the mechanism responsible 236
for the attraction to infested trees). 237
238
Chemical analysis through GC-MS 239
To further elucidate the mechanism underlying the attraction to infested trees, we 240
analyzed the volatile organic compounds emitted by 32 individual trees (16 infested and 241
16 uninfested) right after the behavioral tests, between 16:30-19:00 during 8 242
experimental days. We collected the volatiles of a subset of the total number of trees 243
that were used in the trials with birds (two infested and two uninfested trees each day). 244
We selected one branch of each tree with a similar number of leaves among trees and 245
introduced 20 cm of the branch into a 25x38 cm polyethylene oven bag (Toppits®, 246
Melitta, Lokeren, Belgium). To remove volatile organic compounds, the bags had been 247
heated for 4 hours at 120 ºC before use (Stewart-Jones & Poppy 2006). Bags were 248
fastened to the bark of the branch with tape and one of the two outermost bag corners 249
was cut to allow the placement of a tube containing a steel trap filled with 150 mg 250
Tenax TA and 150 mg Carbopack B. The trap was connected to a vacuum pump. 251
Collection flow rates were set to 200 ml/min. After 2 hours, the traps were removed and 252
capped till analysis. We also measured two background VOC profiles from empty bags 253
on two of the days. The values of compounds in these background samples were 254
12
subtracted from values in the tree samples. Traps were stored at 4 ºC for 10-11 weeks 255
until analysis. Volatiles were desorbed from the traps using an automated 256
thermodesorption unit (model Unity, Markes, Llantrisant, United Kingdom) at 200 °C 257
for 12 min (He flow 30 ml/min) and focused on a cold Tenax trap (–10 °C). After 1 258
minute of dry purging, trapped volatiles were introduced into the GC-MS (model Trace, 259
ThermoFinnigan, Austin, Texas) by heating the cold trap for 3 min to 270 °C. Split ratio 260
was set to 1:4 and the column used was a 30 m x 0.32 mm ID RTX-5 Silms, film 261
thickness 0.33 μm. Temperature program: from 40 °C to 95 °C at 3°C/min. then to 165 262
°C at 2 °C/min, and finally to 250 °C at 15 °C/ min. The volatiles were detected by the 263
MS operating at 70 eV in EI mode. Mass spectra were acquired in full scan mode (33-264
300 AMU. 0.4 scan/sec). Compounds were identified by their mass spectra using 265
deconvolution software (AMDIS) in combination with Nist 98 and Wiley 7th edition 266
spectral libraries and by comparing their linear retention indices. Additionally, mass 267
spectra and/or linear retention indices of chromatographic peaks were compared with 268
values reported in the literature. Additional confirmation for compound identification 269
was obtained by interpolating retention indices of homologous series, or by comparing 270
analytical data with those of reference substances. The integrated signals generated by 271
the AMDIS software from the MS-chromatograms were used for comparison between 272
the treatments. Peak areas in each sample were divided by the total volume in ml that 273
was sampled over the trap, to correct for small differences in flow rates over individual 274
traps. 275
We used an in-house written routine for Orthogonal PLS-DA for MATLAB 276
(Bylesjö et al. 2006), R2010a (Mathworks, Natick MA) to generate a model to describe 277
the general effect of treatment, by contrasting the chemical profiles emitted by trees in 278
the control group against those emitted by trees infested with lepidopteran larvae. The 279
13
data was log-transformed and the average emission of all trees per day was removed 280
from the data, to remove day-to-day variation caused by non-experimental factors. We 281
then determined the two latent variables for the model, by single cross-validation 282
(Westerhuis et al. 2008). We subsequently quantified the significance of the model 283
result by calculating the F-ratio of the obtained class predictions against those from a 284
permutation analysis, where factor ‘time’ was left intact but the ‘treatment’ factor was 285
permuted (Anderson & Ter Braak 2003). These showed that 1000 permuted models all 286
discriminated both treatments less well than that on the original data. We identified the 287
volatiles with largest OPLS-DA weights as significant for the treatment. This analysis 288
did not identify a significant change in the chemistry that underlies this difference 289
during the 14 days of the experiment. 290
291
Coloration measurements 292
We also collected five leaves from infested (N = 16) and uninfested trees (N = 15) used 293
in the behavioral trials and measured coloration. Color measurements were performed 294
by using a USB-2000 spectrophotometer with a DH-2000 deuterium– halogen light 295
source (both Avantes, Eerbeek, The Netherlands). During the measurement of each leaf, 296
we took three replicate readings and obtained the reflectance spectra of each 297
measurement. We calculated the total reflectance of leaves between 300 and 700 nm, 298
which include the spectral range visible to birds (320-700 nm, Cuthill 2006). We also 299
calculated the UV reflectance (between 300-400 nm) and human visible reflectance 300
(400-700 nm). Leaf color measurements were highly repeatable within leaves 301
(repeatability Total reflectance= 0.999; F154, 310 = 6.46; repeatability UV reflectance= 302
0.999; F154, 310 = 3.30; repeatability human-visible reflectance= 0.999; F154, 310 = 6.66) 303
and within trees (repeatability Total reflectance= 0.996; F30, 434 = 4.48; repeatability UV 304
14
reflectance= 0.998; F30, 434 = 2.35; repeatability human-visible reflectance= 0.996; F30, 305
434 = 4.61). Differences between infested and uninfested trees in the reflectance of 306
leaves were analyzed by using GLM with STATISTICA, controlling for the day as a 307
random factor. 308
309
RESULTS 310
Significantly more birds paid the first visit to the infested tree than to the control 311
uninfested tree (Fig. 2a). This preference for the infested tree was found in all 312
treatments (no differences in strength of preference among treatments (χ2 = 0.36, df = 2, 313
P = 0.83; Fig. 1a; significance levels for when the infested tree contained larvae and 314
damaged leaves: Z = 3.14, P = 0.002; only damaged leaves without larvae: Z = 3.08, P = 315
0.002; neither larvae or damaged leaves (“previously damaged”): Z = 3.40, P = 0.0007). 316
The birds also visited the infested tree more frequently than the uninfested tree during 317
the 30 min observation (Fig. 2b). Again, this was similar for all three treatments (χ2 = 318
2.67, df = 2, P = 0.26; “caterpillar”: Z = 3.77, P = 0.0001, “damaged”: Z = 4.01, P < 319
0.0001, and “previously damaged”: Z = 3.80, P = 0.0001), and thus the birds were 320
attracted even when they could not see the caterpillars or their feeding damage. 321
Furthermore, in tests addressing the cues used by the birds, their preference for 322
infested trees, measured as the proportion of visits, was only exhibited when the only 323
cues available were chemical cues (Fig. 3b; Z = 2.99, P = 0.003), but not when there 324
were only visual cues (Z = 0.77, P = 0.44; difference between chemical and visual cues 325
only: χ2 = 5.54, df = 1, P = 0.02). In contrast to the first experiment, in this experiment 326
the first choice did not differ between infested and uninfested trees (P > 0.29 in all 327
cases) or between treatments (χ2 = 0.12, df = 2, P = 0.94, Fig. 3a). 328
15
Leaves from infested trees differed visually from leaves from uninfested trees 329
(Fig. 4), with infested trees having a lower leaf reflectance than uninfested trees both in 330
the visual (F1,22= 9.32, P =0.006) and UV spectral range (F1,22= 4.80, P = 0.04). Trees 331
infested by lepidopteran larvae also differed from uninfested trees in their volatile 332
profiles (see table S1 in Supporting Information), as demonstrated by a validated Partial 333
Least Squares-Discriminant Analysis. They emitted more α-farnesene and dodecanal, 334
while they emitted less 1,2,4 trimethyl benzene, 1-octen-3-ol, methoxy phenyl oxime, 1-335
nonene, and 3-octanol compared to control uninfested trees (Fig. 5). 336
337
DISCUSSION 338
Our results show that great tits exploit herbivore-induced plant volatiles to locate 339
herbivorous prey. The birds were attracted to trees infested by lepidopteran larvae, even 340
when we had removed the larvae and their feeding damage just before the experiment. 341
This allowed us to exclude the option that birds could see the larvae or their feeding 342
damage (Fig. 2). Thus, the preference for infested trees was not due to the visible 343
damage resulting from larval feeding on the leaves, nor by chemical cues associated 344
with the larvae such as silk or feces. A potential explanation for this is that these cues 345
may not accurately signal the current availability of prey in a tree. For example, the 346
presence of damaged leaves on a tree may cause an overestimation of the presence of 347
prey because the damaged leaves remain much longer on the tree than the larvae, which 348
could have been preyed upon or could have left the tree for pupation. 349
Our results show that birds can discriminate between infested and uninfested 350
trees based on the induced response of the tree. Our results are in accordance with 351
previous studies that recorded the attraction of passerine birds to infested trees (Mäntylä 352
et al. 2008a,b). In these previous studies, however, the cues responsible for the 353
16
attraction to infested trees were not separated and it was therefore not possible to 354
conclude whether visual cues, chemical cues or both were responsible for this attraction 355
(Mäntylä et al. 2008a,b). In contrast, we offered birds chemical or visual cues alone or 356
in combination. By doing so, we have shown that bird attraction to infested trees was 357
mainly mediated by chemical cues from the tree, i.e. bird preference for infested trees 358
was still exhibited when the only cues available were plant volatiles (Fig. 3), but not 359
when there were only visual cues. This demonstrates that birds were attracted by the 360
induced emission of volatiles by the tree rather than by the larvae themselves, the visual 361
damage caused by the larvae or the visual cues of undamaged leaves from the infested 362
tree. Similar findings have been reported in previous studies with predatory and 363
parasitoid arthropods (Dicke et al. 1990a; Turlings et al. 1990; Turlings & Tumlinson 364
1992; Vet & Dicke 1992; Mumm & Dicke 2010) but never for vertebrate predators. 365
Infested trees differed visually from uninfested trees (Fig. 4), with infested trees 366
having a lower leaf reflectance than uninfested trees both in the visual and the UV 367
spectral range. Therefore, the coloration of leaves could be a cue to ascertain the level 368
of herbivory of trees. However, visual cues may not be a reliable cue because the 369
reflectance of leaves may be related to other factors affecting trees rather than 370
herbivory, such as sunlight exposure (Mäntylä et al. 2008a). 371
Trees infested by lepidopteran larvae also differed from uninfested trees in their 372
volatile profiles (see table S1), emitting, among others, more α-farnesene compared to 373
control uninfested trees (Fig. 5). The sesquiterpenoid α-farnesene is present both in the 374
headspace of apple leaves (Takabayashi et al. 1991) and apple fruits (Boeve et al. 1996; 375
Landolt et al. 2000), and, at least for fruits, it is involved in the attraction of both 376
herbivorous and predatory insects (Boeve et al. 1996; Landolt et al. 2000). We show 377
that this compound is also present in the headspace of apple trees that are infested with 378
17
winter moth larvae and, thus that the birds can potentially make use of it when locating 379
infested trees. However, further research is needed to establish which compound or 380
mixture of compounds (Bruce & Pickett 2011) is responsible for bird attraction, as well 381
as to understand how differences in emission rates between infested and uninfested trees 382
modulate bird choice behavior. 383
The observation that birds use the volatiles from infested trees to find their prey 384
is in line with other studies on avian olfaction in foraging, and indicates that the 385
importance of olfaction in avian life history may be greater than was previously 386
thought. Phytoplankton releases chemicals to the seawater in response to zooplankton 387
grazing that is converted into dimethyl sulphide (DMS) that is emitted to the air 388
(Pohnert et al. 2007). Hence, DMS signals areas of high productivity in the ocean 389
(Nevitt 2011) and several species of Procellariiformes seabirds (Nevitt et al. 1995; 390
Nevitt 2011) and penguins (Amo et al. 2013) use DMS to locate these productive areas 391
(Nevitt 2011). Indeed, to use chemical cues during foraging seems to be an ancient trait 392
in birds (e.g. Kiwis Apteryx australis (Cunningham et al. 2009); Cathartes vultures 393
(Gomez et al. 1994)), and it persists in several modern lineages (Procellariiforms 394
(Nevitt et al. 1995); chinstrap penguins (Amo et al. 2013); zebra finches (Kelly & 395
Marples 2004); and domestic chicken (Marples & Roper 1996)). 396
The ability to detect the chemical cues emitted by infested trees may especially 397
be important for insectivorous birds such as great tits or blue tits that feed nestlings on 398
lepidopteran larvae, a resource that is variable in space and time (Perrins 1991) and is 399
abundant only for a very short period (Naef-Daenzer et al. 2000). Therefore, the 400
benefits for the birds of using induced volatiles from trees are obvious in terms of 401
increased fitness. From the infested tree’s point of view, the attraction of insectivorous 402
birds can greatly reduce the number of feeding larvae (Mols & Visser 2002), may be 403
18
beneficial in terms of decreased leaf damage and plant mortality (Mäntylä et al. 2011), 404
and therefore, may have a positive impact on fitness. Therefore, our results add to the 405
abundant literature showing that induced plant volatiles attract predators (reviewed by 406
Vet & Dicke 1992; Mumm & Dicke 2010), this time for vertebrate predators. This 407
novel evidence of the ability of insectivorous birds to use chemical cues of infested 408
plants to locate herbivorous prey is exciting because of the high predation rates of birds 409
compared to those of predatory arthropods. This further supports the incentive for plant 410
breeding to enhance the genetic trait underlying the induced volatile emission from 411
plants that are being attacked by insects (Dicke et al. 1990b) and in such a way 412
maximize the impact of insectivorous birds in the biological control of insect pests. 413
414
ACKNOWLEDGEMENTS 415
We thank Piet Drent for allowing us to use the birds for this study and for his helpful 416
comments. We thank Dr. Cornelis A. Hordijk for the GC–MS analyses of volatiles. We 417
are very grateful for their advice to Margriet van Asch and Leonard Holleman on the 418
rearing of winter moth caterpillars, and to Gregor Disveld on the care of apple trees. We 419
thank Marylou Aaldering, Floor Petit, and Janneke Venhorst for their help with the care 420
of great tits. LA was supported by the MEC postdoctoral program and by the Juan de la 421
Cierva program while writing, NMvD and MD by the Earth and Life Sciences Council 422
of the Netherlands Organisation for Scientific Research in the context of the European 423
Science Foundation EUROCORES Programme EuroVOL, and MEV by an NWO-VICI 424
grant. 425
426
427
428
19
429 REFERENCES 430
1. 431
Amo, L., Rodríguez-Gironés, M.A. & Barbosa, A. (2013). Olfactory detection of 432
dimethyl sulphide in a krill-eating Antarctic penguin. Mar. Ecol. Prog. Ser., 474, 433
277–285. 434
2. 435
Anderson, M.J. & Ter Braak, C.J.F. (2003). Permutation tests for multi-factorial 436
analysis of variance. J. Stat. Comput. Simul., 73, 85–113. 437
3. 438
Boeve, J.L., Lengwiler, U., Tollsten, L., Dorn, S. & Turlings, T.C.J. (1996). Volatiles 439
emitted by apple fruitlets infested by larvae of the European apple sawfly. 440
Phytochemistry, 42, 373–381. 441
4. 442
Bruce, T.J.A. & Pickett, J.A. 2011. Perception of plant volatile blends by herbivorous 443
insects – Finding the right mix. Phytochemistry, 72, 1605–1611. 444
5. 445
Bylesjö, M., Rantalainen, M., Cloarec, O., Nicholson, J.K., Holmes, E. & Trygg, J. 446
(2006). OPLS discriminant analysis: combining the strengths of PLS-DA and 447
SIMCA classification. J. Chemometr., 20, 341–351. 448
6. 449
Cunningham, S.J., Castro, I. & Potter, M.A. (2009). The relative importance of 450
olfaction and remote touch in prey detection by North Island brow kiwis. Anim. 451
Behav., 78, 899–905. 452
7. 453
20
Cuthill, I.C. (2006). Color perception. In: Bird Coloration. Volume I: mechanisms and 454
measurements. (eds Hill G.E. McGraw K.J. Harvard University Press, 455
Cambridge, pp. 3–40. 456
8. 457
Dicke, M. & Baldwin, I.T. (2010). The evolutionary context for herbivore-induced plant 458
volatiles: beyond the ‘cry for help’. Trends Plant Sci., 15, 167–175. 459
9. 460
Dicke, M., Van Beek, T.A., Posthumus,M.A., Ben Dom, N., Van Bokhoven, H. & De 461
Groot, A.E. (1990a). Isolation and identification of volatile kairomone that 462
affects acarine predator-prey interactions. Involvement of host plant in its 463
production. J. Chem. Ecol., 16, 381–396. 464
10. 465
Dicke, M., Sabelis, M.W., Takabayashi, J., Bruin, J. & Posthumus, M.A. (1990b). Plant 466
strategies of manipulating predator-prey interactions through allelochemicals: 467
prospects for application in pest control. J. Chem. Ecol., 16, 3091–118. 468
11. 469
Fritzsche Hoballah, M.E. & Turlings, T.C.J. (2001). Experimental evidence that plants 470
under caterpillar attack may benefit from attracting parasitoids. Evol. Ecol. Res., 471
3, 553–565. 472
12. 473
Gomez, L.G., Houston, D.C., Cotton, P. & Tye, A. (1994). The role of greater yellow-474
headed vultures Cathartes melambrotus as scavengers in neotropical forest. Ibis, 475
136, 193–196. 476
13. 477
21
Kelly, D.J. & Marples, N.M. (2004). The effects of novel odour and colour cues on food 478
acceptance by the zebra finch, Taeniopygia guttata. Anim. Behav., 68, 1049–479
1054. 480
14. 481
Landolt, P.J., Brumley, J.A., Smithhisler, C.L., Biddick, L.L. & Hofstetter, R.W. 482
(2000). Apple fruit infested with codling moth are more attractive to neonate 483
codling moth larvae and possess increased amounts of (E,E)-α-farnesene. J. 484
Chem. Ecol., 26, 1685–1699. 485
15. 486
Mäntylä, E., Klemola, T. & Haukioja, E. (2004). Attraction of willow warblers to 487
sawfly-damaged mountain birches: novel function of inducible plant defences? 488
Ecol. Let., 7, 915–918. 489
16. 490
Mäntylä, E., Klemola, T., Sirkiä, P. & Laaksonen, T. (2008a). Low light reflectance 491
may explain the attraction of birds to defoliated trees. Behav. Ecol., 19, 325–492
330. 493
17. 494
Mäntylä, E., Alessio, G.A., Blande, J.D., Heijari, J., Holopainen, J.K., Laaksonen, T., et 495
al. (2008b). From plants to birds: higher avian predation rates in trees 496
responding to insect herbivory. PLoS ONE, 3(7), e2832. 497
18. 498
Mäntylä, E., Klemola, T. & Laaksonen, T. (2011). Birds help plants: a meta–analysis of 499
top–down trophic cascades caused by avian predators. Oecologia, 165, 143–151. 500
19. 501
22
Marples, N.M. & Roper, T.J. (1996). Effects of novel colour and smell on the response 502
of naïve chicks towards food and water. Anim. Behav., 51, 1417–1424. 503
20. 504
Marquis, R.J. & Whelan, C.J. (1994). Insectivorous birds increase growth of white oak 505
through consumption of leaf-chewing insects. Ecology, 75, 2007–2014. 506
21. 507
Mols, C.M.M. & Visser, M.E. (2002). Great tits can reduce caterpillar damage in apple 508
orchards. J. Appl. Ecol., 39, 888–899. 509
22. 510
Mumm, R. & Dicke, M. (2010). Variation in natural plant products and the attraction of 511
bodyguards involved in indirect plant defense. Can. J. Zool., 88, 628–667. 512
23. 513
Naef-Daenzer, L., Naef-Daenzer, B. & Nager, R.G. (2000). Prey selection and foraging 514
performance of breeding Great Tits Parus major in relation to food availability. 515
J. Avian Biol., 31, 206–214. 516
24. 517
Nevitt, G.A. (2011). The neuroecology of dimethyl sulfide: a global-climate regulator 518
turned marine infochemical. Int. Comp. Biol., 51, 819–825. 519
25. 520
Nevitt, G.A., Veit, R.R. & Kareiva, P. (1995). Dimethyl sulphide as a foraging cue for 521
Antarctic procellariiform seabirds. Nature, 376, 680–682. 522
26. 523
Perrins, C.M. (1991). Tits and their caterpillar food supply. Ibis, S1, 133, 49–54. 524
27. 525
23
Pohnert, G., Steinke, M. & Tollrian, R. (2007). Chemical cues, defence metabolites and 526
the shaping of pelagic interspecific interactions. Trends Ecol. Evol., 22, 198–527
204. 528
28. 529
R Development Core Team (2012). R: A Language and Environment for Statistical 530
Computing. R Foundation for Statistical Computing, Vienna. 531
29. 532
Rowland, H.M., Cuthill, I.C., Harvey, I.F., Speed, M.P. & Ruxton, G.D. (2008). Can´t 533
tell the caterpillars from the trees: countershading enhances survival in a 534
woodland. Proc. R. Soc. B, 275, 2539–2545. 535
30. 536
Schoonhoven, L.M., van Loon, J.J.A. & Dicke, M. (2005). Insect-Plant Biology. 537
Oxford. Oxford Univ. Press. 538
31. 539
Schuman, M.C., Barthel, K. & Baldwin, I.T. (2012). Herbivory-induced volatiles 540
function as defenses increasing fitness of the native plant Nicotiana attenuata in 541
nature. eLife, 1, e00007. 542
32. 543
Sipura, M. (1999). Tritrophic interactions: willows, herbivorous insects and 544
insectivorous birds. Oecologia, 121, 537–545. 545
33. 546
Stewart-Jones, A. & Poppy, G.M. (2006). Comparison of glass vessels and plastic bags 547
for enclosing living plant parts for headspace analysis. J. Chem. Ecol., 32, 845–548
864. 549
34. 550
24
Takabayashi, J., Dicke, M. & Posthumus, M.A. (1991). Variation in composition of 551
predator attracting allelochemical emitted by herbivore-infested plants: relative 552
influence of plant and herbivore. Chemoecology, 2, 1–6. 553
35. 554
Turlings, T.C.J. & Tumlinson, J.H. (1992). Systemic release of chemical signals by 555
herbivore-injured corn. Proc. Natl. Acad. Sci. USA, 89, 8399–402. 556
36. 557
Turlings, T.C.J., Tumlinson, J.H. & Lewis, W.J. (1990). Exploitation of herbivore-558
induced plant odors by host-seeking parasitic wasps. Science, 250, 1251–1253. 559
37. 560
Vallat, A., Gu, H. &. Dorn, S. (2005). How rainfall relative humidity and temperature 561
influence volatile emissions from apple trees in situ. Phytochemistry, 66, 1540–562
1550. 563
38. 564
Van Bael, S.A., Brawn, J.D. & Robinson, S.K. (2003). Birds defend trees from 565
herbivores in a Neotropical forest Canopy. PNAS, 100, 8304–8307. 566
39. 567
van Loon, J.J.A., De Boer, J.G. & Dicke, M. (2000). Parasitoid-plant mutualism: 568
parasitoid attack of herbivore increases plant reproduction. Entomol. Exp. Appl., 569
97, 219–227. 570
40. 571
Vet, L.E.M. & Dicke, M. (1992). Ecology of infochemical use by natural enemies in a 572
tritrophic context. Annu. Rev. Entomol., 37, 141–172. 573
41. 574
25
Westerhuis, J. Hoefsloot, H.J., Smit, S., Vis, D., Smilde, A., Velzen, E.J., et al. (2008). 575
Assessment of PLSDA cross validation. Metabolomics, 4, 81–89. 576
577
578
26
579 Figures 580
Figure 1 Schematic representation of the aviaries used for the experiments. Numbers 581
indicate size in meters. This experiment was aimed to disentangle whether birds 582
detected the chemical or the visual cues of infested trees. Each pair of apple trees was 583
located inside a compartment with 2 compartments. The door of one compartment was 584
made of methacrylate to allow the bird to see the tree, but not to smell the tree. The door 585
of the other compartment was made with cotton material to allow the bird to smell the 586
tree, but not to see it. In experiments aimed to examined the attraction of birds to 587
infested trees, the same aviaries were used but without compartments. In these 588
experiments one tree was located in the same place that each compartment separated 589
from the rest of the aviary by a mesh. 590
591
Figure 2 Mean + SE of (a), Number of birds that paid the first visit and (b), Proportion 592
of visits to the experimental infested tree by great tits, Parus major (N = 38), when 593
released in an aviary with two apple trees: one control (uninfested) and one 594
experimental (infested). The experimental, caterpillar-infested tree had one of the 595
treatments: 1) tree with Operopthera brumata caterpillars feeding on the leaves 596
(‘caterpillar’), 2) tree with leaves damaged by caterpillars that were removed before 597
testing (‘damaged leaves’), 3) tree previously damaged by caterpillars but damaged 598
parts and caterpillars were removed before testing (‘previously infested”). 599
600
Figure 3 Mean + SE of (a), Number of birds that paid the first visit and (b), Proportion 601
of visits to the experimental infested tree by great tits, Parus major (N = 35), when 602
released in an aviary with two pairs of apple trees, one control (uninfested) and one 603
27
experimental (infested). The experimental tree pair had one of the following treatments: 604
1) Chemical cues, 2) Visual cues, and 3) Chemical and Visual cues (c.f. “Previously 605
infested” in Fig. 1) released by apple trees under Operopthera brumata caterpillar 606
herbivory. Caterpillars and damaged leaves were removed before the experiment. 607
608
Figure 4 Spectral analysis of leave coloration. Log-transformed mean + SE of (a), Total 609
Reflectance (300-700 nm), (b), human Visible Reflectance (400-700 nm), and (c), UV 610
Reflectance (300-400 nm) of control uninfested apple trees and apples trees infested 611
with Operopthera brumata caterpillars. 612
613
Figure 5 Chemical analysis. Log-transformed mean + SE relative emission rates (peak 614
area per ml volume sampled) of chemical compounds for which emission rates differed, 615
accordingly to PLS-DA, between control uninfested apple trees and apples trees infested 616
with Operopthera brumata caterpillars. 617
618
619
28
620
Fig. 1 621
622
623
624
29
625 Fig. 2 626
(a) 627
628
(b) 629
630
631
30
Fig. 3 632
(a) 633
634
(b) 635
636
31
637
Fig. 4 638
639
32
Fig. 5 640
641 642
33
643 Supporting Information 644
Table S1 Chemical measurements. Mean relative emission rates (peak area per ml 645
volume sampled) + SE of volatiles produced by uninfested trees (N = 16) and trees 646
infested with 30 Operopthera brumata larvae during 3 days (N = 16). Lri = linear 647
retention index. The identification of compounds marked with a * are confirmed by the 648
injection of reference compounds. 649
Compound Lri Uninfested trees Infested trees
2-pentenal 749 658 + 291 15195 + 780
acetic acid butyl ester 812 2567 + 691 7351 + 2649
1,3-octadiene 819 8748 + 7387 524 + 237
ethyl cyclohexane 827 97 + 48 143 + 50
furfural 828 143 + 52 263 + 133
2-hexen-1-ol* 863 3820 + 1558 25894 + 22155
1-hexanol* 873 11814 + 3977 30510 + 21791
1,2-dimethyl benzene 889 1848 + 651 955 + 497
1-nonene 890 23434 + 7987 11360 + 5974
propanoic acid butyl ester 909 2494 + 672 6636 + 2701
methoxy phenyl oxime 910 1354 + 525 637 + 372
tricyclene 916 39 + 22 63 + 20
α–pinene 927 7396+ 2353 3895 + 1022
gamma-valerolactone 949 477 + 154 101 + 42
benzaldehyde 952 5633 + 3431 9997 + 5135
1-heptanol 968 4675 + 2481 1849 + 758
benzene derivate 974 193 + 68 487 + 165
1-octen-3-ol 978 16438 + 13856 1691 + 736
1,5-octadien-3-ol 981 437 + 215 646 + 282
C3 benzene 988 2064 + 612 1019 + 407
2-pentyl furan 990 8037 + 3386 1731 + 833
6-methyl- 5-hepten-2-ol 993 1195 + 410 2157 + 916
3-octanol 996 749 + 358 223 + 143
butanoic acid butyl ester 997 425 + 142 1396 + 529
2-(2-ethoxyethoxy)-ethanol 1004 4947 + 2681 1722 + 591
2-hexen-1-ol-acetate 1018 4466 + 1538 18262 + 13487
2,6-dimethyl nonane 1021 829 + 244 402 + 207
benzene derivate 1022 858 + 478 265 + 84
3-cyclohexen-1-ol-acetate 1029 6761 + 2442 11129 + 4218
34
2-ethyl-1-hexanol 1030 1428 + 803 4039 + 1393
benzyl alcohol 1032 230007 + 119913 121326 + 76838
2-methyl decane 1064 148 + 67 245 + 99
1-octanol 1071 21166 + 14497 4220 + 1818
C4-benzene 1072 122 + 42 59 + 27
methyl-benzoate 1090 594 + 224 255 + 97
4-nonenal 1095 2811 + 769 5112 + 2710
linalool 1099 96415 + 32543 167870 + 68320
benzene derivate 1111 61 + 22 116 + 38
2,6 dimethyl-1,3,5,7 octatetraene 1129 1017 + 381 1730 + 745
nopinone 1132 1336 + 483 724 + 319
benzene acetonitrile 1136 5082 + 2049 7636 + 2237
Z-3-hexenyl iso-butyrate 1148 1452 + 456 2906 + 1496
1-nonanol 1177 8124 + 4631 4647 + 1615
2-nonen-1-ol 1181 329 + 175 624 + 250
octanoic acid 1181 99 + 90 421 + 176
Z-3-hexenyl butyrate 1190 16990 + 5362 36891 + 14122
methyl salicylate* 1191 40180 + 20135 109821 + 52781
branched alkane 1266 250 + 104 528 + 174
1-decanol 1272 3049 + 1698 936 + 437
Z-hexenyl angelate 1280 995 + 363 1646 + 624
Z-3-hexenyl pentanoate 1285 281 + 97 369 + 145
Indole* 1287 8254 + 3492 15220 + 5934
naphthalene derivate 1299 133 + 41 70 + 21
branched alkane 1377 1240 + 474 663 + 214
bourbonene isomer 1380 404174 + 127253 205299 + 69867
β-cubebene 1387 2579 + 1151 1340 + 617
Z-jasmone 1392 6755 + 2499 18163 + 8969
dodecanal 1409 2304 + 929 4190 + 1161
copaene isomer 1432 821 + 392 446 + 203
allo-aromadenderene 1456 9235 + 4382 4235 + 1934
Z-cadina-1(6),4-diene 1461 11500 + 5360 6155 + 3130
D- germacrene 1477 320048 + 158792 95010 + 35948
1-pentadecene 1492 600 + 328 902 + 314
α–farnesene isomer 1 1495 184997 + 138483 51250 + 15825
α–farnesene isomer 2 1506 110191 + 40677 168574 + 69510
benzophenone 1620 57 + 37 170 + 80
isopropyl dodecanoate 1629 1006 + 453 4730 + 2195
1-tetradecanol 1675 210 + 81 418 + 189
cyclohexane undecyl 1763 12 + 9 58 + 29
benzyl-benzoate 1766 1708 + 781 3767 + 2980
35
hexadecanal 1819 723 + 255 1110 + 491
1-hexadecanol 1882 2605 + 2263 4122 + 2206 650 651 652
top related