- Email: [email protected]
The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience
Accepted Manuscript The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience Maryam Yazdani, Michael Keller PII: DOI: Reference:
S1049-9644(15)00097-3 http://dx.doi.org/10.1016/j.biocontrol.2015.05.004 YBCON 3271
To appear in:
Biological Control
Received Date: Accepted Date:
26 December 2014 11 May 2015
Please cite this article as: Yazdani, M., Keller, M., The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience, Biological Control (2015), doi: http:// dx.doi.org/10.1016/j.biocontrol.2015.05.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
For publication in: Biological control The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience Maryam Yazdani1, Michael Keller1* 1
School of Agriculture, Food and Wine, Waite Campus, University of Adelaide, SA 5005
*
Corresponding author: [email protected]
1
1
ABSTRACT
2
Dolichogenidea tasmanica (Hymenoptera: Braconidae) is the most commonly collected
3
parasitoid of light brown apple moth, Epiphyas postvittana (Lepidoptera: Tortricidae;
4
LBAM) in Australia. We studied the functional response of D. tasmanica, and the effect of
5
recent experience on this behaviour. The functional response was evaluated in wind tunnels
6
and enclosed cages. In both arenas, D. tasmanica exhibited a sigmoid functional response,
7
but there was no clear tendency for a deceleration in the functional response curve at high
8
host densities as would be expected with a typical type III functional response. Another
9
experiment revealed that recent experience with high host densities increases the proportion
10
of hosts parasitized by D. tasmanica, which explains much of the difference between the
11
observed functional response curve and a typical type III curve. In general the searching
12
behaviour of D. tasmanica varies in response to host density in a manner that directly affects
13
its searching rate. Our results have contributed to understanding the behaviour of this
14
parasitoid and indicate its capacity to control its host under laboratory conditions. At lower
15
host densities that are characteristic of field populations, D. tasmanica responded in a
16
density-dependent manner that should contribute to suppression of pest populations before
17
they reach economically damaging levels.
18
Key words: density dependence, searching behaviour, vineyard, braconid wasp, parasitoid-
19
host
20 21 22 23 24 25
2
26
Introduction
27
The functional response of a parasitoid to changing host density provides important
28
information on mechanisms underlying parasitoid-host dynamics (Lipcius and Hines, 1986)
29
and is an essential component of parasitoid-host models (Jeschke et al., 2002). The nature of
30
the functional response determines whether a parasitoid is able to regulate the density of its
31
host (Murdoch and Oaten, 1975). Usually, it is classified into one of three general types
32
(Holling, 1959) named I, II and III, which respectively describe curves that are linear,
33
concave increasing to an asymptote and sigmoid when numbers of parasitised hosts per
34
female are plotted against host density (Figure 1a). However, theoretically there are other
35
possible forms, such as type IV, type V and a functional response with parasitoid interference
36
(Hassell, 1978; Abrams, 1982; Taylor, 1984; Turchin, 2001). Parasitoids displaying a type II
37
response cause maximum parasitism rates at low host density. But with type III functional
38
responses, Parasitism rates increase when host density is low, and then decline at higher host
39
densities as handling time reduces the time available for searching. Holling (1959) suggested
40
that the type II response may be typical of invertebrates, including parasitoids, whereas type
41
III responses are characteristic of species that switch between prey species and display
42
learning. Later work indicated that parasitoids can display type III curves (Fernández-Arhex
43
and Corley, 2003). The population consequences of each type of response are different.
44
Whereas a type I response implies a density-independent parasitoid attack rate, a type II
45
response leads to inverse density-dependent predation or parasitism. The type III functional
46
response is the only response which may lead to direct density dependence when host
47
densities are low, and thus can potentially stabilize parasitoid-host interactions (Hassell et al.,
48
1977; Hassell, 1978; Collins et al., 1981; Chesson and Rosenzweig, 1991; Berryman, 1999;
49
Bernstein, 2000; Fernández-arhex and Corley, 2003).
50
3
51
In unstructured models, sigmoid (type III) functional responses have the potential to stabilize
52
parasitoid-host dynamics due to density-dependent mortality at low host densities. In
53
contrast, a type II functional response destabilizes the dynamics because the parasitoids cause
54
an inverse density-dependent mortality of the host (e.g. Murdoch and Oaten, 1975; Hassell,
55
1978). So, distinguishing between type II and type III functional responses is critical in
56
understanding the parasitoid-host dynamics. Due to the importance of functional responses in
57
ecological processes, numerous empirical studies have characterized functional responses in a
58
variety of predator-prey systems (Fernández-arhex Arhex and Corley, 2003; Okuyama,
59
2013). Therefore in order to understand and predict the parasitoid’ success as control agent,
60
the other aspects of the parasitoid behaviour that might affect the functional response deserve
61
more attention (Fernández-arhex and Corley, 2003).
62
Since Holling’s (1959) seminal work, a number of experiments in a variety of species, as
63
well as theoretical studies have been carried out that draw attention to problems involved in
64
measuring the functional response. On the one hand, it has been debated whether the design
65
of some controlled experiments are representative of the behaviour that leads to the true
66
shape of the functional response curve and how these should be carried out. On the other
67
hand, the statistical analyses of the data and the mathematical models used in analyses have
68
been widely discussed (Livdahl and Stiven, 1983; Houck and Strauss, 1985; Williams and
69
Juliano, 1985; Juliano and Williams, 1987; Trexler et al., 1988; Casas and Hulliger, 1994;
70
Manly and Jamienson, 1999; Juliano, 2001; Fernández-arhex and Corley, 2003). Van
71
Lenteren and Bakker (1976) suggested that the apparent absence of a stabilizing density
72
dependence functional response in invertebrate predators or parasitoids may be caused by
73
experimental procedures in which the numbers of prey or hosts at low densities are higher
74
than what can be expected in the field. Hassell et al. (1977) in turn, argued that the practice of
4
75
doing experiments in a relatively simple laboratory universe may ignore the full range of
76
behaviours which invertebrate predators are capable of showing.
77
Juliano and Williams (1987) pointed out that the type of functional response can be most
78
readily determined by analysing the relationship between the proportion of hosts parasitised
79
and host density (Figure 1b).
80
differ fundamentally. The proportion of parasitised hosts is not affected by varying low host
81
densities with a type I response, and with a type II response the proportion declines
82
monotonically from the maximum at the lowest host density. In contrast, the proportion of
83
parasitised hosts the proportion of parasitised hosts increases at low host densities, peaks at
84
an intermediate density and then declines in a manner similar to type I and II responses at
85
high densities.
86
detected using logistic regression (Juliano 2001).
At low host densities, the tree types of functional responses
This difference between type I, II and III responses can be analytically
87
Insect parasitoids are important subjects of behavioural and population studies because
88
they are remarkably common in nature, are typically easy to rear and handle and, more
89
importantly, are key species for the biological control of many insect pests (Waage and
90
Hassell, 1982; Godfray, 1994; Fernández-arhex and Corley, 2003). It is for this reason that
91
functional responses have been investigated in many insect parasitoids (Fernández-arhex and
92
Corley, 2003). Understanding of the nature of the functional response should indicate likely
93
patterns of parasitism, which is important in evaluating population dynamics and the capacity
94
of a parasitoid to contribute to biological control.
95
In this paper we report the results of a series of experiments designed to investigate the
96
functional
response
of
a
parasitic
wasp,
97
(Hymenoptera: Braconidae). It parasitises the light brown apple moth (LBAM), Epiphyas
98
postvittana (Walker) (Lepidoptera: Tortricidae) and other tortricids. LBAM is a polyphagous
99
native species in South-eastern Australia, where it is a key pest in vineyards. It has been 5
Dolichogenidea
tasmanica
(Cameron)
100
introduced to Western Australia, New Zealand, Hawaii, England, and California (Suckling
101
and Brockerhoff, 2010). D. tasmanica is a commonly collected parasitoid of E. postvittana
102
(Paull and Austin, 2006). It is an arrhenotokous, solitary, koinobiont endoparasitoid of the
103
first three instars of LBAM (Yazdani et al., 2014), however no previous study of its
104
functional response has been reported. Our goals were: (1) to characterise the functional
105
response of female D. tasmanica to changing densities of second instar LBAM, (2) to
106
elucidate some of the key factors that affect the shape of functional response curve. A series
107
of experiments was carried out in small wind tunnels to present conditions in which the
108
parasitoid could detect and respond to host cues more naturally because of air flow.
109
Materials and methods
110
Rearing parasitoid and host
111
A laboratory colony of E. postvittana was reared at 22 ± 2º C and a photoperiod of 12 L: 12
112
D on an artificial diet. A culture of D. tasmanica was established from individuals collected
113
from South Australian vineyards. The wasps were reared on larval LBAM infesting plantain,
114
Plantago lanceolata (L.), at 23 ± 2 º C, 14 L: 10 D (for details see Yazdani et al., 2014).
115
Functional response in wind tunnels
116
The functional response of D. tasmanica was investigated in four identical wind tunnels (for
117
details see Yazdani et al., 2014). The wind tunnels had inside dimensions of 35 cm (H)× 50
118
cm (L)× 30 cm (W). The mean wind speed was 29 ± 0.67 cm/s (mean ± SD). Each wind
119
tunnel contained 20 small grape leaves (variety Chardonnay; 3.5 - 4.5 cm L and 4 - 4.5 cm
120
W). Each leaf was placed in a 10 mm diam. × 50 mm glass vial filled with water, and vials
121
were placed 5 cm apart in four rows in the wind tunnel. Six densities of second instar LBAM
122
were tested independently 1, 2, 4, 8, 16 and 32, with 27, 17, 8, 9, 8 and 4 replications
123
respectively. The number of replicates was varied with density in order to achieve equivalent
124
levels of precision across the range of densities that was evaluated. Four experiments were 6
125
run concurrently, with densities chosen at random. For each density, leaves were randomly
126
infested with larvae 24 h before the experiment. Every morning newly emerged females D.
127
tasmanica were collected and caged overnight with 5 males to ensure mating. Naïve 1-2 old
128
mated females were used in the experiments. In order to stimulate the naïve wasps before
129
starting the experiment, each wasp was exposed to a second instar host and allowed to sting it
130
once. The wasp was then released in the wind tunnel 10 cm downwind from the first row of
131
leaves. After 2 h, the wasps were removed, and the LBAM larvae were collected and placed
132
in 100 ml plastic cups that contained a grape leaf for food. They were kept at room
133
temperature for 4 days and then dissected to determine the frequency of parasitism of larvae
134
by D. tasmanica.
135
Functional response in cages
136
In order to determine if the experimental arena affects the shape of the functional response
137
curve, a second experiment was conducted in cages. In this experiment, six densities of
138
second instar LBAM were presented to wasps, 1, 2, 4, 8, 16 and 32, with 26, 14, 6, 6, 10 and
139
4 replications, respectively. The experiments were conducted in plastic containers with inside
140
dimensions of 17 cm (H) × 20 cm (L) × 13 cm (W). The container was modified by removing
141
one side and replacing it with nylon mesh of the same dimension to allow for aeration. For
142
each density, larval LBAM were placed randomly on 6 grape leaves. Each leaf was placed in
143
a 10 mm diam. × 50 mm glass vial filled with water, and vials were placed 5 cm apart in
144
three rows in the cage. The larval LBAM were exposed to a naïve 1-2 old mated female for 2
145
hours. Parasitism data were recorded as described in the previous experiment.
146
Analysis of functional response curves
147
The data from both functional response experiments were analysed using the approach
148
described by Juliano (2001) which comprised two steps: model selection and hypothesis
149
testing. In Julianos’ approach to determine the shape of functional response a logistic 7
150
regression of proportion of parasitised hosts versus number of host recommended. Also,
151
Trexler et al. (1988) demonstrated that the most effective way to distinguish type II and III
152
functional responses involves logistic regression. The second step hypothesis testing involves
153
using nonlinear least-squares regression of number of parasitised host versus number offered
154
to estimate parameters of functional response and to compare parameters of different
155
functional responses. So in our analysing first, the fraction of hosts parasitised vs. number
156
present was subjected to logistic regression with linear, quadratic and cubic terms using the
157
glm function (generalised linear model; family = binomial) in the statistical package R
158
(version 3.1.0 (2014-04-10), "Spring Dance"). In both cases the coefficient of the linear term
159
was found to be positive, which indicates a type III functional response. Therefore the data
160
were then fitted to a type III functional response curve using the nls function (nonlinear least
161
squares) of R, using the equation (1) suggested by Hassell et al. (1977) and elaborated by
162
Juliano (2001):
163
(1)
(
N par = N 1 − e
(
− ( d + bN )Tt 1+ cN + dhN + bhN 2
)
)
164
The mean fraction of parasitised hosts was used in these analyses, since the raw data had a
165
binomial error distribution and as a result had high levels of inherent variation, particularly at
166
low host densities. Nevertheless, the nonlinear regressions did not converge to stable
167
solutions for either data set using either raw data or mean.
168
Effect of experience on type of functional response
169
In the third experiment, we sought to determine if a rewarding experience of foraging for
170
hosts would lead to greater subsequent success in locating hosts compared to wasps that
171
searched in such an arena where no host were present, and hence they had a non-rewarding
172
experience. In each replicate of this experiment, a pair of 1-2 day-old mated females was
173
selected. One of them was designated as having a rewarding experience. It was released for 1 8
174
h in a wind tunnel containing 10 grape leaves, each infested with one second instar LBAM.
175
The other wasp was designated as having a non-rewarding experience. It was released into a
176
wind tunnel containing 10 uninfested grape leaves. The leaves were spaced in two rows
177
across the width of the wind tunnel, and separated by a distance of 5 cm. Immediately after
178
capturing them, the wasps with two types of experience were released separately in two wind
179
tunnels which contained two second instar LBAM that were randomly placed on 20 grape
180
leaves. Host larval positions were the same in both wind tunnels within a replicate. The
181
leaves were arranged in four rows of five leaves and separated by a distance of 5 cm. After 2
182
h exposure to wasps, the infested leaves were collected, placed in plastic cups with a grape
183
leaf and kept at room temperature for 4 days. The larvae were dissected to determine the
184
frequency of parasitism. This experiment was replicated 10 times. The differences in
185
proportions of larvae parasitism between the two types of experience were analysed using
186
Fishers’s two-tailed exact test (Zar, 1984) on pooled numbers.
187
In the fourth experiment, we sought to determine if rewarding and non-rewarding
188
searching experiences have a longer term effect on searching behaviour. This experiment
189
compared the behaviour of rewarding and non-rewarding experience wasps when foraging at
190
three host densities after an interval of 8 h had elapsed following the first bout of searching.
191
In all trials, wasps were released into wind tunnels containing 20 grape leaves using the
192
same methods and arrangement as were used in the functional response experiment to allow
193
the results to be compared. Naïve 1-2 day-old mated female parasitoids were used. Females
194
considered to have a rewarding experience were released for 1 h in a wind tunnel in which
195
each leaf was infested with one second instar LBAM. Females with a non-rewarding
196
experience were released for 1 h in a wind tunnel with uninfested grape leaves. After 1 h, the
197
wasps were transferred to an 18 mm diam. × 50 mm glass vials with a drop of honey and
198
sealed with damp cotton. After 8 h, wasps were released again into wind tunnels that 9
199
contained 4, 8 or 16 second instar LBAM that were randomly placed on 20 grape leaves.
200
Pairs of wasps with rewarding and non-rewarding experience were released into separate
201
wind tunnels with the same density of larval LBAM. After 1 h, wasps were removed and the
202
larvae were placed in plastic cups, held for 4 days with grape leaves and dissected as
203
described previously. This experiment was replicated 8 times for each density and the order
204
of treatments was randomised. The data were analysed with Logistic regression, with linear
205
and quadratic terms for host number using the statistical package R.
206
Results
207
Functional responses in wind tunnels and cages
208
A significant positive linear term in the logistic regression analyses indicated that the data
209
conform best to the type III functional response in both the wind tunnel and cage arenas
210
(Tables 1 and 2; Figures 2 a and b). Parasitism levels were higher at the lower host densities
211
in the cages compared to the wind tunnels, but this trend was reversed at the two highest
212
densities. However in neither case could the data be fitted to the nonlinear type III functional
213
response equation (1) elaborated by Juliano (2001), as the nonlinear least squares analyses
214
did not converge on parameter values that were statistically significant.
215
Effect of experience on searching behaviour
216
Wasp searching behaviour was affected by experience. Females that had a rewarding
217
experience parasitised more larvae at low density than those that had a non-rewarding
218
experience (Table 2). When the interval between initial searching was extended to 8 h,
219
females that had a rewarding experience consistently parasitised more hosts than those that
220
had a non-rewarding experience (Z = -7.715, P < 10-13; Figure 3). Host density also affected
221
the fraction of hosts that were parasitised in a non-linear manner (quadratic term for host
222
number from Logistic regression: Z = -2.185, P = 0.0289), which is consistent with a type III
223
functional response. 10
224
Discussion
225
Invertebrate functional responses are normally measured in small and simple arenas. But
226
Hassell et al. (1977) argued that sigmoid responses are more likely to be found when species
227
can express the full range of searching behavior. In order to diminish the influence of
228
artificial laboratory conditions on behavior, we carried out a series of experiments in wind
229
tunnels to present conditions in which the parasitoid could detect host cues more naturally
230
because of air flow. Also, in our experiments host larvae were distributed randomly among
231
host plant leaves to mimic heterogeneity a wasp would encounter in the field. Thus, the
232
parasitoids could move freely from leaf to leaf and express their full set of host finding
233
behaviours. We also conducted experiments in a simple cage to allow us to determine if the
234
type of arena would lead to a substantial change in the functional response. The cages were
235
smaller than the wind tunnels and the range of densities tested was the same in both arenas,
236
so host density should be perceived to be relatively higher in the cages. Overall the shape of
237
the functional response curve was similar in both arenas (Figures 2 a and b). However,
238
parasitism was higher at the lowest host numbers tested in the cages, while it was higher in
239
the wind tunnels at relatively higher host numbers. These results suggest that wasps could
240
find hosts more easily when the density was low in the smaller area. But in the wind tunnel,
241
the ability to track odour plumes using anemotaxis is likely to have led to greater searching
242
efficiency at higher host densities.
243
Our results in both wind tunnel and cage experiments clearly showed the characteristics
244
of a type III functional response for D. tasmanica at low host densities (Figures 4 a and b).
245
Hassell et al. (1977) argued against the notion that type II functional responses are typical of
246
parasitoids, and suggested that sigmoid type III responses may be common. In subsequent
247
research, hymenoptera species such as Venturia canescens (Grav.) and Campoletis chlorideae
248
(Uchida) (Ichneumonidae), Aphidius uzbekistanicus (Luzhetzki), Diaeretiella rapae 11
249
(M'Intosh) and Aphidius salicis (Haliday) (Aphidiidae), Aphidius colemani (Viereck)
250
(Braconidae), Ibalia leucospoides (Hochenwarth) (Ibaliidae) have been shown to exhibit
251
sigmoid functional responses (Fernández-arhex and Corley, 2003). Fernández-arhex and
252
Corley (2003) suggested that the analyses of some functional response experiments may be
253
overestimating type II curves.
254
experiments may force a type II curve on the insect’s behaviour (van Lenteren and Bakker,
255
1976; Walde and Murdoch, 1988; Ives et al., 1999). Furthermore, type II models may have
256
been used to fit data that could be better served by type III models, especially in older work
257
(Fernández-arhex and Corley, 2003).
For instance, it has been suggested that time-limited
258
It is clear that even the mathematical type III functional response curve does not fit the
259
data (Figures 2 a and b). We could not statistically fit a type III model to the data, even
260
though a Logistic regression analysis clearly showed it is type III in nature. This indicates
261
that the proportion parasitized varied in a complex manner. On the one hand, the results
262
indicate the proportion parasitized varied in a manner consistent with a hyperbolic
263
relationship at low host densities (Juliano, 2001). But at densities of 10 or more hosts per 20
264
leaves, the rate of searching must have increased with increasing host density, which lead to
265
the curve to display characteristics of a type I functional response (Figure 4 a). If the
266
proportion parasitized varied in a purely hyperbolic manner with increasing host density, then
267
the proportion parasitized would asymptotically approach a fixed maximum and the
268
relationship between number parasitised and host density would produce a typical sigmoid
269
curve (Figure 4 b). Thus it seems that a more detailed understanding of the factors that
270
influence proportion parasitized is needed in order to develop models of functional response.
271
We suspected that the lack of fit of the data to a type III model was not simply a
272
statistical problem, so we investigated whether the wasp would change its searching
273
behaviour, and hence the estimated searching rate, at higher densities within the timeframe of 12
274
a two hour experiment. There was a highly significant difference between the parasitism rates
275
associated with wasps with rewarding vs. non-rewarding experience (Figure 3). It seems that
276
the searching behaviour of D. tasmanica varies in response to varying host density in a
277
manner that directly affects its searching rate, even over short periods of time. The results of
278
the experiments reported here elucidate two aspects of the behaviour of D. tasmanica. When
279
wasps search in a non-rewarding area, they subsequently reduce the intensity of their
280
searching activity. But if they search where host densities are high and have a rewarding
281
experience, they subsequently search more intensively which leads to increasing the rates of
282
parasitism.
283
It is noteworthy that the effects of rewarding and non-rewarding experience are observed
284
after the relatively short time of one hour (Table 3), and that these effects persist for at least 8
285
hours (Figure 3). This suggests that the wasp assesses host density and learns the
286
characteristics of a rewarding environment. It implies that D. tasmanica uses information
287
from previously visited patches to adjust its subsequent searching decisions. Learning is
288
regarded as an important factor that leads to the expression of a type III functional response
289
(Real 1979). Parasitoids can change their behaviour in a repeatable way and learn through
290
experience (Vet and Groenewold, 1990; Turlings et al., 1993). Here we conclude that the
291
shape of the observed functional response curve is determined in large part by both host
292
density and the effects of recent experience.
293
Hassell et al. (1977) discussed the implications of type III functional responses for
294
species that search when prey densities are low. A reduction in searching effort in a non-
295
rewarding environment which yields a very low fitness return may be an advantage,
296
particularly if it is not possible to leave that environment. By reducing searching effort, a
297
predator or parasitoid may reduce energetic costs until conditions improve. A similar
298
argument may apply to D. tasmanica searching in a large relatively homogeneous vineyard. 13
299
Although costs and gains are less easily defined for parasitoids, they may involve, for
300
example, the costs that arise from metabolism of carbohydrates (Rivero and Casas, 1999) or
301
exposure of the parasitoid to its own natural enemies, both of which must be balanced against
302
the gains that accrue from the number of hosts successfully parasitised. The sigmoid
303
functional response curve of D. tasmanica at low host densities may reflect a strategy that
304
balances the cost of foraging against expected oviposition success.
305
It is important to put the densities used in our experiments into the context of densities
306
that occur in the field. Supperparasitism was not observed in any dissected larvae, which is
307
consistent with our previous observations (Yazdani et al., 2015). No systematic study of the
308
population dynamics of larval LBAM on grapevines has been published. However, treatment
309
thresholds that guide grape growers in decision-making on the application of insecticides to
310
control damaging infestations of LBAM have been published. On table grapes, the number of
311
larvae per 100 shoots at which insecticidal control is recommended is 10 before flowering
312
and five or less thereafter (Department of Primary Industries Victoria, 2010). The treatment
313
threshold for LBAM on wine grapes is reported to be 20 larvae on foliage per 100 shoots
314
(CCW Cooperative, 2008). Grapes have an indeterminate pattern of annual growth, but over
315
the growing season, between 8 and 30 leaves per shoot are commonly present when LBAM is
316
most likely to reach damaging levels (Lebon et al., 2004). Thus the treatment thresholds for
317
LBAM, which indicate relatively high and damaging populations, are in the order of 0.007 to
318
0.025 larvae/ leaf when between 8 and 30 leaves are present on shoots. These densities are far
319
below those used in the present study. There are two important implications that follow from
320
this. First, the highest densities used in our experiments are arguably extremely high relative
321
to those found in natural populations. Therefore it is not ecologically important that the upper
322
asymptote of the functional response curve was not estimated in this study. The asymptotic
323
maximum percentage parasitism is likely to be ecologically significant only in those species 14
324
where it is approached at commonly observed high densities. Second, the sigmoid shape of
325
the functional response curve at low host densities has the greatest relevance for natural
326
populations (Hassell, 1978; Fernández-arhex and Corley, 2003). This sigmoid shape can lead
327
to direct density-dependent parasitism. Arguably, even lower densities should be used in
328
experiments like ours. But this poses significant practical challenges because larger arenas
329
are needed to determine rates of parasitism at such lower densities, and very high levels of
330
replication are needed to precisely estimate mean parasitism at very low densities. We
331
conclude that experiments on functional response should focus on densities that start at the
332
lowest practical number that can be investigated. Such experiments should be conducted in
333
either outdoor arenas or laboratory arenas like wind tunnels that are conducive to the natural
334
expression of searching behaviour. It is also important to recognise that there are practical
335
limitations to make it impossible to investigate parasitoid behaviour at extremely low host
336
densities with precision.
337
The importance of distinguishing between type II and type III functional responses rests
338
on their very different contributions to population stability (Holling, 1959; Murdoch and
339
Oaten, 1975). Only sigmoid functional responses are density-dependent up to some threshold
340
host density. This contributes to stability if average host densities fall below the threshold.
341
Natural enemies that respond to host in density-dependent manner may be able to quickly
342
suppress pest population before they reach economically damage levels (Cappuccino, 1995;
343
Price, 1997). However, Fernandez-Arhex and Corley (2003) examined the functional
344
responses of parasitoids that have been used in classical biological control programs and
345
found no correlation between the type of response and parasitoid success.
346
Paull, Schellhorn and Austin (2014) conducted large-scale field experiments to quantify
347
and characterize the population response of D. tasmanica to different densities of LBAM in
348
the field. In an apparent contradiction to our results, they concluded that the population 15
349
response of D. tasmanica to varying host density was inversely density-dependent, which
350
implies the species exhibits a type II functional response. However, they did not investigate
351
the components of functional and numerical responses that underlie the pattern of parasitism
352
at the population level. They argued that an inversely density-dependent response may be due
353
to inadequate resources such as access to carbohydrates, specific nutrients, shelter or
354
alternative hosts, which are not available or are in short supply in vineyards. This is because
355
parasitoids are likely to expend more energy and time searching for these resources when
356
they are limiting and, as a result, the time available to maximize their response to increasing
357
host density is reduced (Desouhant et al., 2005; Paull, Schellhorn and Austin, 2014). The
358
realised lifetime fecundity of D. tasmanica is also significantly increased in the presence of
359
flowers, although this is a consequence of the increase in longevity, rather than an increase in
360
daily fecundity (Berndt and Wratten, 2005). And without flowers, offspring sex ratios are
361
strongly male biased, but when females have access to flowers an approximately equal sex
362
ratio is produced. Wasps in our experiment were well-fed, so their behaviour should not have
363
been affected by hunger. We conclude that the functional response must be considered in
364
conjunction with other aspects of biology and behaviour when developing models of
365
parasitoid-host population dynamics.
366
Our results showed that D. tasmanica can parasitise LBAM in a density-dependent
367
manner at low host densities, which is important in regulation of host populations. This
368
suggests that D. tasmanica can contribute to biological control of LBAM. Additional studies
369
are needed, however, to investigate the role that experience and learning play in shaping the
370
functional response over the lifetime of a wasp. It is known that experience over time can
371
influence the searching behaviour of the parasitoid V. canescens (Froissart et al., 2012). It is
372
likely that generalist species like D. tasmanica similarly responds to experience with a range
373
of host-related cues over the span of its adult life. 16
374
Acknowledgments
375
Financial support for this research was provided by the Adelaide International Scholarship
376
(ASI) to Maryam Yazdani. The wind tunnels were constructed by Samantha Scarratt and
377
Michael Keller.
378
References
379
Abrams, P., 1982. Functional responses of optimal foragers. Am. Nat. 120, 382-390.
380
Berndt, L.A., Wratten, S.D., 2005. Effects of alyssum flowers on the longevity, fecundity,
381
and sex ratio of the leafroller parasitoid Dolichogenidea tasmanica. Biol. Control 32,
382
65-69. DOI: 10.1016/j.biocontrol.2004.07.014
383
Bernstein, C., 2000. Host-parasitoid models. The story of successful failure. In: Hochberg
384
ME, Ives AR (eds) Parasitoid Population Biology. Princeton University Press,
385
Princeton, New Jersey, USA, pp. 41-57.
386
Berryman, A.A., 1999. The theoretical foundations of biological control. In: Hawkins BA,
387
Cornell HV (eds) Theoretical Approaches to Biological Control. Cambridge University
388
Press, Cambridge, pp. 3-21.
389
Cappuccino, N., 1995. Novel approaches to the study of population dynamics. In: Cappucino
390
N, Price PW (eds) New Approaches and Synthesis. Academic Press, New York, New
391
York, USA, pp. 3-16.
392 393 394 395
Casas, J., Hulliger, B., 1994. Statistical analysis of functional response experiments. Biocontrol Sci. Techn. 4, 133-145. Chesson, P., Rosenzweig, M., 1991. Behaviour, heterogeneity, and the dynamics of interacting species. Ecology 72, 1187-1195.
396
Collins, M.D., Ward, S.A., Dixon, A.F.G., 1981. Handling time and the functional response
397
of Aphelinus thomsoni, a predator and parasite of the aphid Drepanosiphum
398
platanoidis. J. Anim. Ecol. 50, 479-487. 17
399
Department of Primary Industries, Victoria, 2010. Pest monitoring for table grape exports.
400
Department of Primary Industries, Melbourne, Australia. ISBN 978-1-74264-564-3
401
[www.australiangrapes.com.au/__data/assets/pdf_file/0003/11388/Pest-monitoring-for-
402
table-grape-exports-Protocol.pdf, accessed 7 June 2014]
403
Desouhant, E., Driessen, G., Amat, I., Bernstein, C., 2005. Host and food searching in a
404
parasitic wasp Venturia canescens: a trade-off between current and future
405
reproduction? Anim. Behav. 70, 145-152.
406
Fernández-arhex, V., Corley, J.C., 2003. The functional response of parasitoids and its
407
implications for biological control. Biocontrol Sci. Techn. 13, 403-413. DOI:
408
10.1080/0958315031000104523
409
Froissart, L., Bernstein, C., Humblot, L., Amat, I., Desouhant, E., 2012. Facing multiple
410
information sources while foraging on successive patches: how does a parasitoid deal
411
with experience? Anim. Behav. 83, 189-199.
412 413 414 415 416 417 418 419 420 421 422 423
Godfray, H.C.J., 1994. Parasitoids: Behavioral and Evolutionary Ecology, Princeton University Press, Princeton, New Jersey, USA. Hassell, M.P., Lawton, J.H., Beddington, J.R., 1977. Sigmoid functional response by invertebrate predators and parasitoids. J. Anim. Ecol. 46, 249-262. Hassell, M.P., 1978. The Dynamics of Arthropod Predator-Prey Systems, Princeton University Press, Princeton, New Jersey, USA. Hassell, M.P., 2000. The Spatial and Temporal Dynamics of Host-Parasitoid Interactions. Oxford Series in Ecology and Evolution, Oxford University Press, London. Holling, C.S., 1959. Some characteristics of simple types of predation and parasitism. Can. Entomol. 91, 385-398. Houck, M.A., Strauss, R.E., 1985. The comparative study of functional responses: experimental design and statistical interpretation. Can. Entomol. 117, 617-629. 18
424
Ives, A.R., Schooler, S.S., Jagar, V.J., Knuteson, S.E., Grbic, M., Settle, W.H., 1999.
425
Variability and parasitoid foraging efficiency: a case study of pea aphids and Aphidius
426
ervi. Am. Nat., 154, 652-673.
427
Jeschke, M.J., Kopp, M., Tollrian, R., 2002. Predator functional response: discriminating
428
between handling and digesting prey. Ecol. Monograph 72, 95-112. DOI:
429
10.2307/3100087
430
Juliano, S.A., Williams, F.M., 1987. A comparison of methods for estimating the functional
431
response parameters of the random predator equation. J. Anim. Ecol. 56, 641-653.
432
Juliano, S.A., 2001. Nonlinear Curve Fitting: Predation and Functional Response Curves.
433
Scheiner SM, Gurevitch J (eds.), Design and Analysis of Ecological Experiments
434
Chapman and Hall, New York, USA, pp. 178-196.
435
Lebon, E., Pellegrino, A., Tardieu, F., Lecoeur, J., 2004. Shoot development in grapevine
436
(Vitis vinifera) is affected by the modular branching pattern of the stem and intra- and
437
inter-shoot trophic competition. Ann. of Bot. 93, 263-274. DOI: 10.1093/aob/mch038
438 439 440 441 442 443 444 445 446 447
Lipcius, R.N., Hines, A.H., 1986. Variable function responses of a marine predator in dissimilar homogeneous microhabitats. Ecology 67, 1361-1371. Livdahl, T.P., Stiven, A.E., 1983. Statistical difficulties in the analysis of predator functional response data. Can. Entomol. 115, 1365-1370. Manly, B.F.J., Jamienson, C.D., 1999. Functional response and parallel curve analysis. Oikos. 85, 523-528. DOI: 10.2307/3546701 Murdoch, W.W., Oaten, A., 1975. Predation and population stability. Adv. Ecol. Res. 9, 2131. Okuyama, T., 2013. On selection of functional response models: Holling’s models and more. BioControl 58, 293-298.
19
448
Paull, C.A., Austin, A.D., 2006. The hymenopteran parasitoids of light brown apple moth,
449
Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae) in Australia. Aust. J.
450
Entomol. 45, 142-156. DOI: 10.1111/j.1440-6055.2006.00524.x
451
Paull, C.A., Schellhorn, N.A., Austin, A.D., 2014. Response to host density by the parasitoid
452
Dolichogenidea tasmanica (Hymenoptera: Braconidae) and the influence of grapevine
453
variety. Bull. Entomol. Res. 104, 79-87.
454
DOI: http://dx.doi.org.proxy.library.adelaide.edu.au/10.1017/S0007485313000527
455
CCW Cooperative, Berri, South Australia. 6 p. [www.ccwcoop.com.au/__files/f/4382/CCW
456
FACT SHEET 4 - LBAM.pdf, accessed 7 June 2014]
457
Price, P.W., 1997. Insect Ecology. Academic Press, New York, USA.
458
Real, L.A., 1979. Ecological determinants of functional response. Ecology 60, 481-485.
459
Rivero, A., Casas, J., 1999. Incorporating physiology into parasitoid behavioural ecology: the
460
allocation of nutritional resources. Res. Popul. Ecol. 41, 39-45.
461
Suckling, D.M., Brockerhoff, E.G., 2010. Invasion Biology, Ecology and Management of the
462
Light Brown Apple Moth (Tortricidae). Annu. Rev. Entomol. 55, 285-306.
463
DOI: 10.1146/annurev-ento-112408-085311
464
Taylor, R.J., 1984. Predation. Chapman and Hall, New York, USA.
465
Trexler, J.C., Mccullogh, C.E., Travis, J., 1988. How can the functional response best be
466 467 468
determined? Oecologia 76, 206-214. DIO: 10.1007/BF00378751 Turchin, P., 2001. Complex Population Dynamics: A Theoretical/ Empirical Synthesis. Princeton University Press, Princeton, New Jersey, USA.
469
Turlings, T.C.L., Wäckers, F.L., Vet, L.E.M., Lewis, W.J., Tumlinson, J.H., 1993. Learning
470
of Host-Finding Cues by Hymenopterous Parasitoids. Papaj DR, Lewis AC (eds.)
471
Insect learning: ecology and evolutionary perspectives. Springer Science+Business
472
Media, Dordrecht, Netherlands, pp. 51-78. 20
473 474 475 476 477 478 479 480 481
van Lenteren, J.C., Bakker, K., 1976. Functional responses in invertebrates. Neth. J. Zool. 26, 567-572. Vet, L.E.M., Groenewold, A.W., 1990. Semiochemicals and learning in parasitoids. J. Chem. Ecol. 16, 119-35. Waage, J.K., Hassell, M.P., 1982. Parasitoids as biological control agents: a fundamental approach. Parasitology 84, 241-268. Walde, S.J., Murdoch, W.W., 1988. Spatial density dependence in parasitoids. Annu. Rev. Entomol., 33, 441-466. Williams, F.M., Juliano, S.A., 1985. Further difficulties in the analysis of functional response
482
experiments
483
DOI: http://dx.doi.org/10.4039/Ent117631-5
484
and
a
resolution.
Entomol.
117,
631-640.
Yazdani, M., Feng, Y., Glatz, R., Keller, M.A., 2014. Host stage preference of tasmanica
(Cameron)
485
Dolichogenidea
486
Entomology. DOI: 10.1111/aen.12130
487
Can.
(Hymenoptera:
Braconidae).
Austral
Zar, J.H., 1984. Biostatistical analysis. Prentice-Hall Inc., Englewood Clifffs, New Jersey.
488 489 490 491 492 493 494 495 496 497
21
498
Figure legends
499
Figure1. Three types of functional response.
500
hosts parasitised and host density.
501
parasitised and host density.
502
Figure2. Mean fraction of larvae parasitised (± standard error) by D. tasmanica for the 6
503
densities (1, 2, 4, 6, 8, 16 and 32) of second instar LBAM in the wind tunnels. The dotted line
504
depicts the expected functional response if atype III data and the black line showsthe fitted
505
Logistic regression curve a)in wind tunnels and b) in small cages.
506
Figure 3. The effect of previous rewarding or non-rewarding foraging experience by D.
507
tasmanica on the mean fraction of larvae parasitised (± standard error) when foraging 8 h
508
later at densities of 4, 8 and 16 larvae per 20 grape leaves.
509
Figure . Type III functional response curves (solid lines; b = 0.005, c = 0.04, d = 0.000, h = 8,
510
T = 120) and the effect of a switch to a type I response (dashed lines). a. Relationship
511
between host number and fraction parasitised in the wind tunnel. b. Relationship between
512
host number and number parasitised in the wind tunnel
a) The relationship between the number of
b) The relationship between the proportion of hosts
513 514 515 516 517 518 519 520 521 522
22
523
Figure 1.
524
a)
Number of hosts parasitised/individual
525
Type I Type II Type III
0 0
526
b)
Type I Type II Type III
Proportion of hosts parasitised/individual
527
0 528
0
Host density
529 530 531 532
23
533 534 535
Figure 2
536
a)
Fraction parasitised (mean ± SE)
1
Observed
Type III
0.8 0.6 0.4 0.2 0 0
537
Logistic
5
10
15 20 Host density
538
539
540
541
542
543
544
24
25
30
35
545
b)
Fraction parasitised (mean ± SE)
546
1
Observed
Logistic
Type III
0.8 0.6 0.4 0.2 0 0
547
5
10
15
20
Host number
548 549 550 551 552 553 554 555 556 557 558
25
25
30
35
559 560
Figure 3
Fraction parasitised (mean + SE)
0.8
Rewarding
0.6
0.4
0.2
0 0
561
Non-rewarding
4
8 12 Number of host larvae
562
563
564
565
566 567 568 569 570 571
26
16
20
572
Figure 4
574
a.
Fraction parasitised
573
0.8 0.6 0.4 0.2 0
575
0
10
20 Host number
30
40
0
10
20 Host number
30
40
b.
Number parasitised
25 20 15 10 5 0
576 577 578 579 580
27
581 582
Table 1. Results of logistic regression analysis of the fraction of hosts parasitised by D.
583
tasmanica vs. host number in small wind tunnels.
584
Coefficient
Estimate
Std. Error
z value
Prob.
Intercept
-0.8238131
0.4015574
-2.052
0.04021
Host number
0.4047707
0.1447773
2.796
0.00518
(Host number)2
-0.0279509
0.0116589
-2.397
0.01651
(Host number)3
0.0005274
0.0002362
2.233
0.02553
585
Null deviance: 122.98 on 72 degrees of freedom
586
Residual deviance: 108.59 on 69 degrees of freedom
587 588 589 590 591 592 593 594 595 596 597 598
28
599 600
Table 2. Results of logistic regression analysis of the fraction of hosts parasitised by D.
601
tasmanica vs. host number in cages. Coefficient
Estimate
Std. Error
z value
Prob.
Intercept
-0.0741800
0.4347062
-0.171
0.86450
Host number
0.3851388
0.1675998
2.298
0.02156
(Host number)2
-0.0368871
0.0134945
-2.733
0.00627
(Host number)3
0.0007866
0.0002712
2.900
0.00373
602
Null deviance: 151.49 on 65 degrees of freedom
603
Residual deviance: 136.03 on 62 degrees of freedom
604 605 606 607 608 609 610 611 612 613 614 615 616
29
617 618
Table 3. The effect of experience on the frequency of parasitism of second instar LBAM by
619
D. tasmanica when presented with two hosts in a wind tunnel.
620
Previous experience Fate of larva
Rewarding
Non-rewarding
Unparasitised
5
16
Parasitised
15
4
Total No. wasps
10
10
Fisher’s two-tailed exact test, P = 0.0012
30
1, 2, 4, 8, 16 or 32 larvae Wind
Fan
Fraction parasitised
Variable density of larval Light Brown Apple Moth 0.8 0.6 0.4 0.2
0 0
10
20
30
Host number
Wind tunnel
40
•
Dolichogenidea tasmanica exhibited a sigmoid functional response in wind tunnels and enclosed cages.
•
Parasitism rates were lower in the cages, possibly due to the lack of moving air which provides a directional cue.
•
Recent experience with high host densities increases the searching rate of D. tasmanica,
•
At lower host densities that are characteristic of wild populations, D. tasmanica responded in a density-dependent manner
•
This manner should contribute to suppression of pest populations before they reach economically damage levels.
S1049-9644(15)00097-3 http://dx.doi.org/10.1016/j.biocontrol.2015.05.004 YBCON 3271
To appear in:
Biological Control
Received Date: Accepted Date:
26 December 2014 11 May 2015
Please cite this article as: Yazdani, M., Keller, M., The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience, Biological Control (2015), doi: http:// dx.doi.org/10.1016/j.biocontrol.2015.05.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
For publication in: Biological control The shape of the functional response curve of Dolichogenidea tasmanica (Hymenoptera: Braconidae) is affected by recent experience Maryam Yazdani1, Michael Keller1* 1
School of Agriculture, Food and Wine, Waite Campus, University of Adelaide, SA 5005
*
Corresponding author: [email protected]
1
1
ABSTRACT
2
Dolichogenidea tasmanica (Hymenoptera: Braconidae) is the most commonly collected
3
parasitoid of light brown apple moth, Epiphyas postvittana (Lepidoptera: Tortricidae;
4
LBAM) in Australia. We studied the functional response of D. tasmanica, and the effect of
5
recent experience on this behaviour. The functional response was evaluated in wind tunnels
6
and enclosed cages. In both arenas, D. tasmanica exhibited a sigmoid functional response,
7
but there was no clear tendency for a deceleration in the functional response curve at high
8
host densities as would be expected with a typical type III functional response. Another
9
experiment revealed that recent experience with high host densities increases the proportion
10
of hosts parasitized by D. tasmanica, which explains much of the difference between the
11
observed functional response curve and a typical type III curve. In general the searching
12
behaviour of D. tasmanica varies in response to host density in a manner that directly affects
13
its searching rate. Our results have contributed to understanding the behaviour of this
14
parasitoid and indicate its capacity to control its host under laboratory conditions. At lower
15
host densities that are characteristic of field populations, D. tasmanica responded in a
16
density-dependent manner that should contribute to suppression of pest populations before
17
they reach economically damaging levels.
18
Key words: density dependence, searching behaviour, vineyard, braconid wasp, parasitoid-
19
host
20 21 22 23 24 25
2
26
Introduction
27
The functional response of a parasitoid to changing host density provides important
28
information on mechanisms underlying parasitoid-host dynamics (Lipcius and Hines, 1986)
29
and is an essential component of parasitoid-host models (Jeschke et al., 2002). The nature of
30
the functional response determines whether a parasitoid is able to regulate the density of its
31
host (Murdoch and Oaten, 1975). Usually, it is classified into one of three general types
32
(Holling, 1959) named I, II and III, which respectively describe curves that are linear,
33
concave increasing to an asymptote and sigmoid when numbers of parasitised hosts per
34
female are plotted against host density (Figure 1a). However, theoretically there are other
35
possible forms, such as type IV, type V and a functional response with parasitoid interference
36
(Hassell, 1978; Abrams, 1982; Taylor, 1984; Turchin, 2001). Parasitoids displaying a type II
37
response cause maximum parasitism rates at low host density. But with type III functional
38
responses, Parasitism rates increase when host density is low, and then decline at higher host
39
densities as handling time reduces the time available for searching. Holling (1959) suggested
40
that the type II response may be typical of invertebrates, including parasitoids, whereas type
41
III responses are characteristic of species that switch between prey species and display
42
learning. Later work indicated that parasitoids can display type III curves (Fernández-Arhex
43
and Corley, 2003). The population consequences of each type of response are different.
44
Whereas a type I response implies a density-independent parasitoid attack rate, a type II
45
response leads to inverse density-dependent predation or parasitism. The type III functional
46
response is the only response which may lead to direct density dependence when host
47
densities are low, and thus can potentially stabilize parasitoid-host interactions (Hassell et al.,
48
1977; Hassell, 1978; Collins et al., 1981; Chesson and Rosenzweig, 1991; Berryman, 1999;
49
Bernstein, 2000; Fernández-arhex and Corley, 2003).
50
3
51
In unstructured models, sigmoid (type III) functional responses have the potential to stabilize
52
parasitoid-host dynamics due to density-dependent mortality at low host densities. In
53
contrast, a type II functional response destabilizes the dynamics because the parasitoids cause
54
an inverse density-dependent mortality of the host (e.g. Murdoch and Oaten, 1975; Hassell,
55
1978). So, distinguishing between type II and type III functional responses is critical in
56
understanding the parasitoid-host dynamics. Due to the importance of functional responses in
57
ecological processes, numerous empirical studies have characterized functional responses in a
58
variety of predator-prey systems (Fernández-arhex Arhex and Corley, 2003; Okuyama,
59
2013). Therefore in order to understand and predict the parasitoid’ success as control agent,
60
the other aspects of the parasitoid behaviour that might affect the functional response deserve
61
more attention (Fernández-arhex and Corley, 2003).
62
Since Holling’s (1959) seminal work, a number of experiments in a variety of species, as
63
well as theoretical studies have been carried out that draw attention to problems involved in
64
measuring the functional response. On the one hand, it has been debated whether the design
65
of some controlled experiments are representative of the behaviour that leads to the true
66
shape of the functional response curve and how these should be carried out. On the other
67
hand, the statistical analyses of the data and the mathematical models used in analyses have
68
been widely discussed (Livdahl and Stiven, 1983; Houck and Strauss, 1985; Williams and
69
Juliano, 1985; Juliano and Williams, 1987; Trexler et al., 1988; Casas and Hulliger, 1994;
70
Manly and Jamienson, 1999; Juliano, 2001; Fernández-arhex and Corley, 2003). Van
71
Lenteren and Bakker (1976) suggested that the apparent absence of a stabilizing density
72
dependence functional response in invertebrate predators or parasitoids may be caused by
73
experimental procedures in which the numbers of prey or hosts at low densities are higher
74
than what can be expected in the field. Hassell et al. (1977) in turn, argued that the practice of
4
75
doing experiments in a relatively simple laboratory universe may ignore the full range of
76
behaviours which invertebrate predators are capable of showing.
77
Juliano and Williams (1987) pointed out that the type of functional response can be most
78
readily determined by analysing the relationship between the proportion of hosts parasitised
79
and host density (Figure 1b).
80
differ fundamentally. The proportion of parasitised hosts is not affected by varying low host
81
densities with a type I response, and with a type II response the proportion declines
82
monotonically from the maximum at the lowest host density. In contrast, the proportion of
83
parasitised hosts the proportion of parasitised hosts increases at low host densities, peaks at
84
an intermediate density and then declines in a manner similar to type I and II responses at
85
high densities.
86
detected using logistic regression (Juliano 2001).
At low host densities, the tree types of functional responses
This difference between type I, II and III responses can be analytically
87
Insect parasitoids are important subjects of behavioural and population studies because
88
they are remarkably common in nature, are typically easy to rear and handle and, more
89
importantly, are key species for the biological control of many insect pests (Waage and
90
Hassell, 1982; Godfray, 1994; Fernández-arhex and Corley, 2003). It is for this reason that
91
functional responses have been investigated in many insect parasitoids (Fernández-arhex and
92
Corley, 2003). Understanding of the nature of the functional response should indicate likely
93
patterns of parasitism, which is important in evaluating population dynamics and the capacity
94
of a parasitoid to contribute to biological control.
95
In this paper we report the results of a series of experiments designed to investigate the
96
functional
response
of
a
parasitic
wasp,
97
(Hymenoptera: Braconidae). It parasitises the light brown apple moth (LBAM), Epiphyas
98
postvittana (Walker) (Lepidoptera: Tortricidae) and other tortricids. LBAM is a polyphagous
99
native species in South-eastern Australia, where it is a key pest in vineyards. It has been 5
Dolichogenidea
tasmanica
(Cameron)
100
introduced to Western Australia, New Zealand, Hawaii, England, and California (Suckling
101
and Brockerhoff, 2010). D. tasmanica is a commonly collected parasitoid of E. postvittana
102
(Paull and Austin, 2006). It is an arrhenotokous, solitary, koinobiont endoparasitoid of the
103
first three instars of LBAM (Yazdani et al., 2014), however no previous study of its
104
functional response has been reported. Our goals were: (1) to characterise the functional
105
response of female D. tasmanica to changing densities of second instar LBAM, (2) to
106
elucidate some of the key factors that affect the shape of functional response curve. A series
107
of experiments was carried out in small wind tunnels to present conditions in which the
108
parasitoid could detect and respond to host cues more naturally because of air flow.
109
Materials and methods
110
Rearing parasitoid and host
111
A laboratory colony of E. postvittana was reared at 22 ± 2º C and a photoperiod of 12 L: 12
112
D on an artificial diet. A culture of D. tasmanica was established from individuals collected
113
from South Australian vineyards. The wasps were reared on larval LBAM infesting plantain,
114
Plantago lanceolata (L.), at 23 ± 2 º C, 14 L: 10 D (for details see Yazdani et al., 2014).
115
Functional response in wind tunnels
116
The functional response of D. tasmanica was investigated in four identical wind tunnels (for
117
details see Yazdani et al., 2014). The wind tunnels had inside dimensions of 35 cm (H)× 50
118
cm (L)× 30 cm (W). The mean wind speed was 29 ± 0.67 cm/s (mean ± SD). Each wind
119
tunnel contained 20 small grape leaves (variety Chardonnay; 3.5 - 4.5 cm L and 4 - 4.5 cm
120
W). Each leaf was placed in a 10 mm diam. × 50 mm glass vial filled with water, and vials
121
were placed 5 cm apart in four rows in the wind tunnel. Six densities of second instar LBAM
122
were tested independently 1, 2, 4, 8, 16 and 32, with 27, 17, 8, 9, 8 and 4 replications
123
respectively. The number of replicates was varied with density in order to achieve equivalent
124
levels of precision across the range of densities that was evaluated. Four experiments were 6
125
run concurrently, with densities chosen at random. For each density, leaves were randomly
126
infested with larvae 24 h before the experiment. Every morning newly emerged females D.
127
tasmanica were collected and caged overnight with 5 males to ensure mating. Naïve 1-2 old
128
mated females were used in the experiments. In order to stimulate the naïve wasps before
129
starting the experiment, each wasp was exposed to a second instar host and allowed to sting it
130
once. The wasp was then released in the wind tunnel 10 cm downwind from the first row of
131
leaves. After 2 h, the wasps were removed, and the LBAM larvae were collected and placed
132
in 100 ml plastic cups that contained a grape leaf for food. They were kept at room
133
temperature for 4 days and then dissected to determine the frequency of parasitism of larvae
134
by D. tasmanica.
135
Functional response in cages
136
In order to determine if the experimental arena affects the shape of the functional response
137
curve, a second experiment was conducted in cages. In this experiment, six densities of
138
second instar LBAM were presented to wasps, 1, 2, 4, 8, 16 and 32, with 26, 14, 6, 6, 10 and
139
4 replications, respectively. The experiments were conducted in plastic containers with inside
140
dimensions of 17 cm (H) × 20 cm (L) × 13 cm (W). The container was modified by removing
141
one side and replacing it with nylon mesh of the same dimension to allow for aeration. For
142
each density, larval LBAM were placed randomly on 6 grape leaves. Each leaf was placed in
143
a 10 mm diam. × 50 mm glass vial filled with water, and vials were placed 5 cm apart in
144
three rows in the cage. The larval LBAM were exposed to a naïve 1-2 old mated female for 2
145
hours. Parasitism data were recorded as described in the previous experiment.
146
Analysis of functional response curves
147
The data from both functional response experiments were analysed using the approach
148
described by Juliano (2001) which comprised two steps: model selection and hypothesis
149
testing. In Julianos’ approach to determine the shape of functional response a logistic 7
150
regression of proportion of parasitised hosts versus number of host recommended. Also,
151
Trexler et al. (1988) demonstrated that the most effective way to distinguish type II and III
152
functional responses involves logistic regression. The second step hypothesis testing involves
153
using nonlinear least-squares regression of number of parasitised host versus number offered
154
to estimate parameters of functional response and to compare parameters of different
155
functional responses. So in our analysing first, the fraction of hosts parasitised vs. number
156
present was subjected to logistic regression with linear, quadratic and cubic terms using the
157
glm function (generalised linear model; family = binomial) in the statistical package R
158
(version 3.1.0 (2014-04-10), "Spring Dance"). In both cases the coefficient of the linear term
159
was found to be positive, which indicates a type III functional response. Therefore the data
160
were then fitted to a type III functional response curve using the nls function (nonlinear least
161
squares) of R, using the equation (1) suggested by Hassell et al. (1977) and elaborated by
162
Juliano (2001):
163
(1)
(
N par = N 1 − e
(
− ( d + bN )Tt 1+ cN + dhN + bhN 2
)
)
164
The mean fraction of parasitised hosts was used in these analyses, since the raw data had a
165
binomial error distribution and as a result had high levels of inherent variation, particularly at
166
low host densities. Nevertheless, the nonlinear regressions did not converge to stable
167
solutions for either data set using either raw data or mean.
168
Effect of experience on type of functional response
169
In the third experiment, we sought to determine if a rewarding experience of foraging for
170
hosts would lead to greater subsequent success in locating hosts compared to wasps that
171
searched in such an arena where no host were present, and hence they had a non-rewarding
172
experience. In each replicate of this experiment, a pair of 1-2 day-old mated females was
173
selected. One of them was designated as having a rewarding experience. It was released for 1 8
174
h in a wind tunnel containing 10 grape leaves, each infested with one second instar LBAM.
175
The other wasp was designated as having a non-rewarding experience. It was released into a
176
wind tunnel containing 10 uninfested grape leaves. The leaves were spaced in two rows
177
across the width of the wind tunnel, and separated by a distance of 5 cm. Immediately after
178
capturing them, the wasps with two types of experience were released separately in two wind
179
tunnels which contained two second instar LBAM that were randomly placed on 20 grape
180
leaves. Host larval positions were the same in both wind tunnels within a replicate. The
181
leaves were arranged in four rows of five leaves and separated by a distance of 5 cm. After 2
182
h exposure to wasps, the infested leaves were collected, placed in plastic cups with a grape
183
leaf and kept at room temperature for 4 days. The larvae were dissected to determine the
184
frequency of parasitism. This experiment was replicated 10 times. The differences in
185
proportions of larvae parasitism between the two types of experience were analysed using
186
Fishers’s two-tailed exact test (Zar, 1984) on pooled numbers.
187
In the fourth experiment, we sought to determine if rewarding and non-rewarding
188
searching experiences have a longer term effect on searching behaviour. This experiment
189
compared the behaviour of rewarding and non-rewarding experience wasps when foraging at
190
three host densities after an interval of 8 h had elapsed following the first bout of searching.
191
In all trials, wasps were released into wind tunnels containing 20 grape leaves using the
192
same methods and arrangement as were used in the functional response experiment to allow
193
the results to be compared. Naïve 1-2 day-old mated female parasitoids were used. Females
194
considered to have a rewarding experience were released for 1 h in a wind tunnel in which
195
each leaf was infested with one second instar LBAM. Females with a non-rewarding
196
experience were released for 1 h in a wind tunnel with uninfested grape leaves. After 1 h, the
197
wasps were transferred to an 18 mm diam. × 50 mm glass vials with a drop of honey and
198
sealed with damp cotton. After 8 h, wasps were released again into wind tunnels that 9
199
contained 4, 8 or 16 second instar LBAM that were randomly placed on 20 grape leaves.
200
Pairs of wasps with rewarding and non-rewarding experience were released into separate
201
wind tunnels with the same density of larval LBAM. After 1 h, wasps were removed and the
202
larvae were placed in plastic cups, held for 4 days with grape leaves and dissected as
203
described previously. This experiment was replicated 8 times for each density and the order
204
of treatments was randomised. The data were analysed with Logistic regression, with linear
205
and quadratic terms for host number using the statistical package R.
206
Results
207
Functional responses in wind tunnels and cages
208
A significant positive linear term in the logistic regression analyses indicated that the data
209
conform best to the type III functional response in both the wind tunnel and cage arenas
210
(Tables 1 and 2; Figures 2 a and b). Parasitism levels were higher at the lower host densities
211
in the cages compared to the wind tunnels, but this trend was reversed at the two highest
212
densities. However in neither case could the data be fitted to the nonlinear type III functional
213
response equation (1) elaborated by Juliano (2001), as the nonlinear least squares analyses
214
did not converge on parameter values that were statistically significant.
215
Effect of experience on searching behaviour
216
Wasp searching behaviour was affected by experience. Females that had a rewarding
217
experience parasitised more larvae at low density than those that had a non-rewarding
218
experience (Table 2). When the interval between initial searching was extended to 8 h,
219
females that had a rewarding experience consistently parasitised more hosts than those that
220
had a non-rewarding experience (Z = -7.715, P < 10-13; Figure 3). Host density also affected
221
the fraction of hosts that were parasitised in a non-linear manner (quadratic term for host
222
number from Logistic regression: Z = -2.185, P = 0.0289), which is consistent with a type III
223
functional response. 10
224
Discussion
225
Invertebrate functional responses are normally measured in small and simple arenas. But
226
Hassell et al. (1977) argued that sigmoid responses are more likely to be found when species
227
can express the full range of searching behavior. In order to diminish the influence of
228
artificial laboratory conditions on behavior, we carried out a series of experiments in wind
229
tunnels to present conditions in which the parasitoid could detect host cues more naturally
230
because of air flow. Also, in our experiments host larvae were distributed randomly among
231
host plant leaves to mimic heterogeneity a wasp would encounter in the field. Thus, the
232
parasitoids could move freely from leaf to leaf and express their full set of host finding
233
behaviours. We also conducted experiments in a simple cage to allow us to determine if the
234
type of arena would lead to a substantial change in the functional response. The cages were
235
smaller than the wind tunnels and the range of densities tested was the same in both arenas,
236
so host density should be perceived to be relatively higher in the cages. Overall the shape of
237
the functional response curve was similar in both arenas (Figures 2 a and b). However,
238
parasitism was higher at the lowest host numbers tested in the cages, while it was higher in
239
the wind tunnels at relatively higher host numbers. These results suggest that wasps could
240
find hosts more easily when the density was low in the smaller area. But in the wind tunnel,
241
the ability to track odour plumes using anemotaxis is likely to have led to greater searching
242
efficiency at higher host densities.
243
Our results in both wind tunnel and cage experiments clearly showed the characteristics
244
of a type III functional response for D. tasmanica at low host densities (Figures 4 a and b).
245
Hassell et al. (1977) argued against the notion that type II functional responses are typical of
246
parasitoids, and suggested that sigmoid type III responses may be common. In subsequent
247
research, hymenoptera species such as Venturia canescens (Grav.) and Campoletis chlorideae
248
(Uchida) (Ichneumonidae), Aphidius uzbekistanicus (Luzhetzki), Diaeretiella rapae 11
249
(M'Intosh) and Aphidius salicis (Haliday) (Aphidiidae), Aphidius colemani (Viereck)
250
(Braconidae), Ibalia leucospoides (Hochenwarth) (Ibaliidae) have been shown to exhibit
251
sigmoid functional responses (Fernández-arhex and Corley, 2003). Fernández-arhex and
252
Corley (2003) suggested that the analyses of some functional response experiments may be
253
overestimating type II curves.
254
experiments may force a type II curve on the insect’s behaviour (van Lenteren and Bakker,
255
1976; Walde and Murdoch, 1988; Ives et al., 1999). Furthermore, type II models may have
256
been used to fit data that could be better served by type III models, especially in older work
257
(Fernández-arhex and Corley, 2003).
For instance, it has been suggested that time-limited
258
It is clear that even the mathematical type III functional response curve does not fit the
259
data (Figures 2 a and b). We could not statistically fit a type III model to the data, even
260
though a Logistic regression analysis clearly showed it is type III in nature. This indicates
261
that the proportion parasitized varied in a complex manner. On the one hand, the results
262
indicate the proportion parasitized varied in a manner consistent with a hyperbolic
263
relationship at low host densities (Juliano, 2001). But at densities of 10 or more hosts per 20
264
leaves, the rate of searching must have increased with increasing host density, which lead to
265
the curve to display characteristics of a type I functional response (Figure 4 a). If the
266
proportion parasitized varied in a purely hyperbolic manner with increasing host density, then
267
the proportion parasitized would asymptotically approach a fixed maximum and the
268
relationship between number parasitised and host density would produce a typical sigmoid
269
curve (Figure 4 b). Thus it seems that a more detailed understanding of the factors that
270
influence proportion parasitized is needed in order to develop models of functional response.
271
We suspected that the lack of fit of the data to a type III model was not simply a
272
statistical problem, so we investigated whether the wasp would change its searching
273
behaviour, and hence the estimated searching rate, at higher densities within the timeframe of 12
274
a two hour experiment. There was a highly significant difference between the parasitism rates
275
associated with wasps with rewarding vs. non-rewarding experience (Figure 3). It seems that
276
the searching behaviour of D. tasmanica varies in response to varying host density in a
277
manner that directly affects its searching rate, even over short periods of time. The results of
278
the experiments reported here elucidate two aspects of the behaviour of D. tasmanica. When
279
wasps search in a non-rewarding area, they subsequently reduce the intensity of their
280
searching activity. But if they search where host densities are high and have a rewarding
281
experience, they subsequently search more intensively which leads to increasing the rates of
282
parasitism.
283
It is noteworthy that the effects of rewarding and non-rewarding experience are observed
284
after the relatively short time of one hour (Table 3), and that these effects persist for at least 8
285
hours (Figure 3). This suggests that the wasp assesses host density and learns the
286
characteristics of a rewarding environment. It implies that D. tasmanica uses information
287
from previously visited patches to adjust its subsequent searching decisions. Learning is
288
regarded as an important factor that leads to the expression of a type III functional response
289
(Real 1979). Parasitoids can change their behaviour in a repeatable way and learn through
290
experience (Vet and Groenewold, 1990; Turlings et al., 1993). Here we conclude that the
291
shape of the observed functional response curve is determined in large part by both host
292
density and the effects of recent experience.
293
Hassell et al. (1977) discussed the implications of type III functional responses for
294
species that search when prey densities are low. A reduction in searching effort in a non-
295
rewarding environment which yields a very low fitness return may be an advantage,
296
particularly if it is not possible to leave that environment. By reducing searching effort, a
297
predator or parasitoid may reduce energetic costs until conditions improve. A similar
298
argument may apply to D. tasmanica searching in a large relatively homogeneous vineyard. 13
299
Although costs and gains are less easily defined for parasitoids, they may involve, for
300
example, the costs that arise from metabolism of carbohydrates (Rivero and Casas, 1999) or
301
exposure of the parasitoid to its own natural enemies, both of which must be balanced against
302
the gains that accrue from the number of hosts successfully parasitised. The sigmoid
303
functional response curve of D. tasmanica at low host densities may reflect a strategy that
304
balances the cost of foraging against expected oviposition success.
305
It is important to put the densities used in our experiments into the context of densities
306
that occur in the field. Supperparasitism was not observed in any dissected larvae, which is
307
consistent with our previous observations (Yazdani et al., 2015). No systematic study of the
308
population dynamics of larval LBAM on grapevines has been published. However, treatment
309
thresholds that guide grape growers in decision-making on the application of insecticides to
310
control damaging infestations of LBAM have been published. On table grapes, the number of
311
larvae per 100 shoots at which insecticidal control is recommended is 10 before flowering
312
and five or less thereafter (Department of Primary Industries Victoria, 2010). The treatment
313
threshold for LBAM on wine grapes is reported to be 20 larvae on foliage per 100 shoots
314
(CCW Cooperative, 2008). Grapes have an indeterminate pattern of annual growth, but over
315
the growing season, between 8 and 30 leaves per shoot are commonly present when LBAM is
316
most likely to reach damaging levels (Lebon et al., 2004). Thus the treatment thresholds for
317
LBAM, which indicate relatively high and damaging populations, are in the order of 0.007 to
318
0.025 larvae/ leaf when between 8 and 30 leaves are present on shoots. These densities are far
319
below those used in the present study. There are two important implications that follow from
320
this. First, the highest densities used in our experiments are arguably extremely high relative
321
to those found in natural populations. Therefore it is not ecologically important that the upper
322
asymptote of the functional response curve was not estimated in this study. The asymptotic
323
maximum percentage parasitism is likely to be ecologically significant only in those species 14
324
where it is approached at commonly observed high densities. Second, the sigmoid shape of
325
the functional response curve at low host densities has the greatest relevance for natural
326
populations (Hassell, 1978; Fernández-arhex and Corley, 2003). This sigmoid shape can lead
327
to direct density-dependent parasitism. Arguably, even lower densities should be used in
328
experiments like ours. But this poses significant practical challenges because larger arenas
329
are needed to determine rates of parasitism at such lower densities, and very high levels of
330
replication are needed to precisely estimate mean parasitism at very low densities. We
331
conclude that experiments on functional response should focus on densities that start at the
332
lowest practical number that can be investigated. Such experiments should be conducted in
333
either outdoor arenas or laboratory arenas like wind tunnels that are conducive to the natural
334
expression of searching behaviour. It is also important to recognise that there are practical
335
limitations to make it impossible to investigate parasitoid behaviour at extremely low host
336
densities with precision.
337
The importance of distinguishing between type II and type III functional responses rests
338
on their very different contributions to population stability (Holling, 1959; Murdoch and
339
Oaten, 1975). Only sigmoid functional responses are density-dependent up to some threshold
340
host density. This contributes to stability if average host densities fall below the threshold.
341
Natural enemies that respond to host in density-dependent manner may be able to quickly
342
suppress pest population before they reach economically damage levels (Cappuccino, 1995;
343
Price, 1997). However, Fernandez-Arhex and Corley (2003) examined the functional
344
responses of parasitoids that have been used in classical biological control programs and
345
found no correlation between the type of response and parasitoid success.
346
Paull, Schellhorn and Austin (2014) conducted large-scale field experiments to quantify
347
and characterize the population response of D. tasmanica to different densities of LBAM in
348
the field. In an apparent contradiction to our results, they concluded that the population 15
349
response of D. tasmanica to varying host density was inversely density-dependent, which
350
implies the species exhibits a type II functional response. However, they did not investigate
351
the components of functional and numerical responses that underlie the pattern of parasitism
352
at the population level. They argued that an inversely density-dependent response may be due
353
to inadequate resources such as access to carbohydrates, specific nutrients, shelter or
354
alternative hosts, which are not available or are in short supply in vineyards. This is because
355
parasitoids are likely to expend more energy and time searching for these resources when
356
they are limiting and, as a result, the time available to maximize their response to increasing
357
host density is reduced (Desouhant et al., 2005; Paull, Schellhorn and Austin, 2014). The
358
realised lifetime fecundity of D. tasmanica is also significantly increased in the presence of
359
flowers, although this is a consequence of the increase in longevity, rather than an increase in
360
daily fecundity (Berndt and Wratten, 2005). And without flowers, offspring sex ratios are
361
strongly male biased, but when females have access to flowers an approximately equal sex
362
ratio is produced. Wasps in our experiment were well-fed, so their behaviour should not have
363
been affected by hunger. We conclude that the functional response must be considered in
364
conjunction with other aspects of biology and behaviour when developing models of
365
parasitoid-host population dynamics.
366
Our results showed that D. tasmanica can parasitise LBAM in a density-dependent
367
manner at low host densities, which is important in regulation of host populations. This
368
suggests that D. tasmanica can contribute to biological control of LBAM. Additional studies
369
are needed, however, to investigate the role that experience and learning play in shaping the
370
functional response over the lifetime of a wasp. It is known that experience over time can
371
influence the searching behaviour of the parasitoid V. canescens (Froissart et al., 2012). It is
372
likely that generalist species like D. tasmanica similarly responds to experience with a range
373
of host-related cues over the span of its adult life. 16
374
Acknowledgments
375
Financial support for this research was provided by the Adelaide International Scholarship
376
(ASI) to Maryam Yazdani. The wind tunnels were constructed by Samantha Scarratt and
377
Michael Keller.
378
References
379
Abrams, P., 1982. Functional responses of optimal foragers. Am. Nat. 120, 382-390.
380
Berndt, L.A., Wratten, S.D., 2005. Effects of alyssum flowers on the longevity, fecundity,
381
and sex ratio of the leafroller parasitoid Dolichogenidea tasmanica. Biol. Control 32,
382
65-69. DOI: 10.1016/j.biocontrol.2004.07.014
383
Bernstein, C., 2000. Host-parasitoid models. The story of successful failure. In: Hochberg
384
ME, Ives AR (eds) Parasitoid Population Biology. Princeton University Press,
385
Princeton, New Jersey, USA, pp. 41-57.
386
Berryman, A.A., 1999. The theoretical foundations of biological control. In: Hawkins BA,
387
Cornell HV (eds) Theoretical Approaches to Biological Control. Cambridge University
388
Press, Cambridge, pp. 3-21.
389
Cappuccino, N., 1995. Novel approaches to the study of population dynamics. In: Cappucino
390
N, Price PW (eds) New Approaches and Synthesis. Academic Press, New York, New
391
York, USA, pp. 3-16.
392 393 394 395
Casas, J., Hulliger, B., 1994. Statistical analysis of functional response experiments. Biocontrol Sci. Techn. 4, 133-145. Chesson, P., Rosenzweig, M., 1991. Behaviour, heterogeneity, and the dynamics of interacting species. Ecology 72, 1187-1195.
396
Collins, M.D., Ward, S.A., Dixon, A.F.G., 1981. Handling time and the functional response
397
of Aphelinus thomsoni, a predator and parasite of the aphid Drepanosiphum
398
platanoidis. J. Anim. Ecol. 50, 479-487. 17
399
Department of Primary Industries, Victoria, 2010. Pest monitoring for table grape exports.
400
Department of Primary Industries, Melbourne, Australia. ISBN 978-1-74264-564-3
401
[www.australiangrapes.com.au/__data/assets/pdf_file/0003/11388/Pest-monitoring-for-
402
table-grape-exports-Protocol.pdf, accessed 7 June 2014]
403
Desouhant, E., Driessen, G., Amat, I., Bernstein, C., 2005. Host and food searching in a
404
parasitic wasp Venturia canescens: a trade-off between current and future
405
reproduction? Anim. Behav. 70, 145-152.
406
Fernández-arhex, V., Corley, J.C., 2003. The functional response of parasitoids and its
407
implications for biological control. Biocontrol Sci. Techn. 13, 403-413. DOI:
408
10.1080/0958315031000104523
409
Froissart, L., Bernstein, C., Humblot, L., Amat, I., Desouhant, E., 2012. Facing multiple
410
information sources while foraging on successive patches: how does a parasitoid deal
411
with experience? Anim. Behav. 83, 189-199.
412 413 414 415 416 417 418 419 420 421 422 423
Godfray, H.C.J., 1994. Parasitoids: Behavioral and Evolutionary Ecology, Princeton University Press, Princeton, New Jersey, USA. Hassell, M.P., Lawton, J.H., Beddington, J.R., 1977. Sigmoid functional response by invertebrate predators and parasitoids. J. Anim. Ecol. 46, 249-262. Hassell, M.P., 1978. The Dynamics of Arthropod Predator-Prey Systems, Princeton University Press, Princeton, New Jersey, USA. Hassell, M.P., 2000. The Spatial and Temporal Dynamics of Host-Parasitoid Interactions. Oxford Series in Ecology and Evolution, Oxford University Press, London. Holling, C.S., 1959. Some characteristics of simple types of predation and parasitism. Can. Entomol. 91, 385-398. Houck, M.A., Strauss, R.E., 1985. The comparative study of functional responses: experimental design and statistical interpretation. Can. Entomol. 117, 617-629. 18
424
Ives, A.R., Schooler, S.S., Jagar, V.J., Knuteson, S.E., Grbic, M., Settle, W.H., 1999.
425
Variability and parasitoid foraging efficiency: a case study of pea aphids and Aphidius
426
ervi. Am. Nat., 154, 652-673.
427
Jeschke, M.J., Kopp, M., Tollrian, R., 2002. Predator functional response: discriminating
428
between handling and digesting prey. Ecol. Monograph 72, 95-112. DOI:
429
10.2307/3100087
430
Juliano, S.A., Williams, F.M., 1987. A comparison of methods for estimating the functional
431
response parameters of the random predator equation. J. Anim. Ecol. 56, 641-653.
432
Juliano, S.A., 2001. Nonlinear Curve Fitting: Predation and Functional Response Curves.
433
Scheiner SM, Gurevitch J (eds.), Design and Analysis of Ecological Experiments
434
Chapman and Hall, New York, USA, pp. 178-196.
435
Lebon, E., Pellegrino, A., Tardieu, F., Lecoeur, J., 2004. Shoot development in grapevine
436
(Vitis vinifera) is affected by the modular branching pattern of the stem and intra- and
437
inter-shoot trophic competition. Ann. of Bot. 93, 263-274. DOI: 10.1093/aob/mch038
438 439 440 441 442 443 444 445 446 447
Lipcius, R.N., Hines, A.H., 1986. Variable function responses of a marine predator in dissimilar homogeneous microhabitats. Ecology 67, 1361-1371. Livdahl, T.P., Stiven, A.E., 1983. Statistical difficulties in the analysis of predator functional response data. Can. Entomol. 115, 1365-1370. Manly, B.F.J., Jamienson, C.D., 1999. Functional response and parallel curve analysis. Oikos. 85, 523-528. DOI: 10.2307/3546701 Murdoch, W.W., Oaten, A., 1975. Predation and population stability. Adv. Ecol. Res. 9, 2131. Okuyama, T., 2013. On selection of functional response models: Holling’s models and more. BioControl 58, 293-298.
19
448
Paull, C.A., Austin, A.D., 2006. The hymenopteran parasitoids of light brown apple moth,
449
Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae) in Australia. Aust. J.
450
Entomol. 45, 142-156. DOI: 10.1111/j.1440-6055.2006.00524.x
451
Paull, C.A., Schellhorn, N.A., Austin, A.D., 2014. Response to host density by the parasitoid
452
Dolichogenidea tasmanica (Hymenoptera: Braconidae) and the influence of grapevine
453
variety. Bull. Entomol. Res. 104, 79-87.
454
DOI: http://dx.doi.org.proxy.library.adelaide.edu.au/10.1017/S0007485313000527
455
CCW Cooperative, Berri, South Australia. 6 p. [www.ccwcoop.com.au/__files/f/4382/CCW
456
FACT SHEET 4 - LBAM.pdf, accessed 7 June 2014]
457
Price, P.W., 1997. Insect Ecology. Academic Press, New York, USA.
458
Real, L.A., 1979. Ecological determinants of functional response. Ecology 60, 481-485.
459
Rivero, A., Casas, J., 1999. Incorporating physiology into parasitoid behavioural ecology: the
460
allocation of nutritional resources. Res. Popul. Ecol. 41, 39-45.
461
Suckling, D.M., Brockerhoff, E.G., 2010. Invasion Biology, Ecology and Management of the
462
Light Brown Apple Moth (Tortricidae). Annu. Rev. Entomol. 55, 285-306.
463
DOI: 10.1146/annurev-ento-112408-085311
464
Taylor, R.J., 1984. Predation. Chapman and Hall, New York, USA.
465
Trexler, J.C., Mccullogh, C.E., Travis, J., 1988. How can the functional response best be
466 467 468
determined? Oecologia 76, 206-214. DIO: 10.1007/BF00378751 Turchin, P., 2001. Complex Population Dynamics: A Theoretical/ Empirical Synthesis. Princeton University Press, Princeton, New Jersey, USA.
469
Turlings, T.C.L., Wäckers, F.L., Vet, L.E.M., Lewis, W.J., Tumlinson, J.H., 1993. Learning
470
of Host-Finding Cues by Hymenopterous Parasitoids. Papaj DR, Lewis AC (eds.)
471
Insect learning: ecology and evolutionary perspectives. Springer Science+Business
472
Media, Dordrecht, Netherlands, pp. 51-78. 20
473 474 475 476 477 478 479 480 481
van Lenteren, J.C., Bakker, K., 1976. Functional responses in invertebrates. Neth. J. Zool. 26, 567-572. Vet, L.E.M., Groenewold, A.W., 1990. Semiochemicals and learning in parasitoids. J. Chem. Ecol. 16, 119-35. Waage, J.K., Hassell, M.P., 1982. Parasitoids as biological control agents: a fundamental approach. Parasitology 84, 241-268. Walde, S.J., Murdoch, W.W., 1988. Spatial density dependence in parasitoids. Annu. Rev. Entomol., 33, 441-466. Williams, F.M., Juliano, S.A., 1985. Further difficulties in the analysis of functional response
482
experiments
483
DOI: http://dx.doi.org/10.4039/Ent117631-5
484
and
a
resolution.
Entomol.
117,
631-640.
Yazdani, M., Feng, Y., Glatz, R., Keller, M.A., 2014. Host stage preference of tasmanica
(Cameron)
485
Dolichogenidea
486
Entomology. DOI: 10.1111/aen.12130
487
Can.
(Hymenoptera:
Braconidae).
Austral
Zar, J.H., 1984. Biostatistical analysis. Prentice-Hall Inc., Englewood Clifffs, New Jersey.
488 489 490 491 492 493 494 495 496 497
21
498
Figure legends
499
Figure1. Three types of functional response.
500
hosts parasitised and host density.
501
parasitised and host density.
502
Figure2. Mean fraction of larvae parasitised (± standard error) by D. tasmanica for the 6
503
densities (1, 2, 4, 6, 8, 16 and 32) of second instar LBAM in the wind tunnels. The dotted line
504
depicts the expected functional response if atype III data and the black line showsthe fitted
505
Logistic regression curve a)in wind tunnels and b) in small cages.
506
Figure 3. The effect of previous rewarding or non-rewarding foraging experience by D.
507
tasmanica on the mean fraction of larvae parasitised (± standard error) when foraging 8 h
508
later at densities of 4, 8 and 16 larvae per 20 grape leaves.
509
Figure . Type III functional response curves (solid lines; b = 0.005, c = 0.04, d = 0.000, h = 8,
510
T = 120) and the effect of a switch to a type I response (dashed lines). a. Relationship
511
between host number and fraction parasitised in the wind tunnel. b. Relationship between
512
host number and number parasitised in the wind tunnel
a) The relationship between the number of
b) The relationship between the proportion of hosts
513 514 515 516 517 518 519 520 521 522
22
523
Figure 1.
524
a)
Number of hosts parasitised/individual
525
Type I Type II Type III
0 0
526
b)
Type I Type II Type III
Proportion of hosts parasitised/individual
527
0 528
0
Host density
529 530 531 532
23
533 534 535
Figure 2
536
a)
Fraction parasitised (mean ± SE)
1
Observed
Type III
0.8 0.6 0.4 0.2 0 0
537
Logistic
5
10
15 20 Host density
538
539
540
541
542
543
544
24
25
30
35
545
b)
Fraction parasitised (mean ± SE)
546
1
Observed
Logistic
Type III
0.8 0.6 0.4 0.2 0 0
547
5
10
15
20
Host number
548 549 550 551 552 553 554 555 556 557 558
25
25
30
35
559 560
Figure 3
Fraction parasitised (mean + SE)
0.8
Rewarding
0.6
0.4
0.2
0 0
561
Non-rewarding
4
8 12 Number of host larvae
562
563
564
565
566 567 568 569 570 571
26
16
20
572
Figure 4
574
a.
Fraction parasitised
573
0.8 0.6 0.4 0.2 0
575
0
10
20 Host number
30
40
0
10
20 Host number
30
40
b.
Number parasitised
25 20 15 10 5 0
576 577 578 579 580
27
581 582
Table 1. Results of logistic regression analysis of the fraction of hosts parasitised by D.
583
tasmanica vs. host number in small wind tunnels.
584
Coefficient
Estimate
Std. Error
z value
Prob.
Intercept
-0.8238131
0.4015574
-2.052
0.04021
Host number
0.4047707
0.1447773
2.796
0.00518
(Host number)2
-0.0279509
0.0116589
-2.397
0.01651
(Host number)3
0.0005274
0.0002362
2.233
0.02553
585
Null deviance: 122.98 on 72 degrees of freedom
586
Residual deviance: 108.59 on 69 degrees of freedom
587 588 589 590 591 592 593 594 595 596 597 598
28
599 600
Table 2. Results of logistic regression analysis of the fraction of hosts parasitised by D.
601
tasmanica vs. host number in cages. Coefficient
Estimate
Std. Error
z value
Prob.
Intercept
-0.0741800
0.4347062
-0.171
0.86450
Host number
0.3851388
0.1675998
2.298
0.02156
(Host number)2
-0.0368871
0.0134945
-2.733
0.00627
(Host number)3
0.0007866
0.0002712
2.900
0.00373
602
Null deviance: 151.49 on 65 degrees of freedom
603
Residual deviance: 136.03 on 62 degrees of freedom
604 605 606 607 608 609 610 611 612 613 614 615 616
29
617 618
Table 3. The effect of experience on the frequency of parasitism of second instar LBAM by
619
D. tasmanica when presented with two hosts in a wind tunnel.
620
Previous experience Fate of larva
Rewarding
Non-rewarding
Unparasitised
5
16
Parasitised
15
4
Total No. wasps
10
10
Fisher’s two-tailed exact test, P = 0.0012
30
1, 2, 4, 8, 16 or 32 larvae Wind
Fan
Fraction parasitised
Variable density of larval Light Brown Apple Moth 0.8 0.6 0.4 0.2
0 0
10
20
30
Host number
Wind tunnel
40
•
Dolichogenidea tasmanica exhibited a sigmoid functional response in wind tunnels and enclosed cages.
•
Parasitism rates were lower in the cages, possibly due to the lack of moving air which provides a directional cue.
•
Recent experience with high host densities increases the searching rate of D. tasmanica,
•
At lower host densities that are characteristic of wild populations, D. tasmanica responded in a density-dependent manner
•
This manner should contribute to suppression of pest populations before they reach economically damage levels.