Are associational refuges species‐specific?

Abstract

Associational refuges – reduced herbivory on plants in presence of other plant species – may be caused by general and species‐specific plant characteristics. However, the species specificity of associational refuges has rarely been evaluated.
This study examined the species specificity of one known example of associational refuges, the forb Lythrum salicaria and the monophagous insect herbivores Galerucella calmariensis , G . pusilla and Nanophyes marmoratus . The underlying mechanism was examined in order to evaluate connections between mechanisms and species specificity.
Laboratory studies showed that N . marmoratus but not Galerucella individuals were attracted by odour from undamaged host plants, and that neither species was distracted by odour from Myrica gale .
Field experiments showed that three non‐host plant neighbours with an appearance similar to M . gale , and artificial shrubs, reduced the abundance and egg‐laying of Galerucella species by 70–90%. The abundance of N. marmoratus was increased 18‐fold on plants in thickets compared with outside.
. The different responses by N . marmoratus and the Galerucella species to plant neighbours appear to be because N. marmoratus , but not Galerucella , uses olfactory information in the initial host‐finding process.

Full text

Functional
Ecology
2003
17
, 87–93
© 2003 British
Ecological Society
87
Blackwell Science, Ltd
Are associational refuges species-specific?
P. A. HAMBÄCK*†, J. PETTERSSON‡
and
L. ERICSON§
*
Department of Ecology and Crop Production Science, Section for Landscape Ecology and
‡
Department of
Entomology, Swedish University of Agricultural Sciences, PO Box 7044, SE-750 07 Uppsala, Sweden, and
§
Department of Ecology and Environmental Sciences, Umeå University, SE-901 87 Umeå, Sweden
Summary
1.
Associational refuges – reduced herbivory on plants in presence of other plant spe-
cies – may be caused by general and species-specific plant characteristics. However, the
species specificity of associational refuges has rarely been evaluated.
2.
This study examined the species specificity of one known example of associational
refuges, the forb
Lythrum salicaria
and the monophagous insect herbivores
Galerucella
calmariensis
,
G
.
pusilla
and
Nanophyes marmoratus
. The underlying mechanism was
examined in order to evaluate connections between mechanisms and species specificity.
3.
Laboratory studies showed that
N
.
marmoratus
but not
Galerucella
individuals were
attracted by odour from undamaged host plants, and that neither species was dis-
tracted by odour from
Myrica gale
.
4.
Field experiments showed that three non-host plant neighbours with an appearance
similar to
M
.
gale
, and artificial shrubs, reduced the abundance and egg-laying of
Galerucella
species by 70–90%. The abundance of
N. marmoratus
was increased 18-fold
on plants in thickets compared with outside.
5
. The different responses by
N
.
marmoratus
and the
Galerucella
species to plant
neighbours appear to be because
N. marmoratus
, but not
Galerucella
, uses olfactory
information in the initial host-finding process.
Key-words
: Associational refuges,
Galerucella calmariensis
,
Galerucella pusilla
, host-finding interference,
Lythrum salicaria
,
Nanophyes marmoratus
, olfactometer
Functional Ecology
(2003)
17
, 87–93
Introduction
Plant productivity increases with plant diversity in
many plant communities (Hector
et al
. 1999; Tilman,
Wedin & Knops 1996), and one mechanism that may
account for this positive correlation is reduced her-
bivore damage on plants in species-rich habitats
(Andow 1991; Koricheva
et al
. 2000; Mulder
et al
. 1999;
Tonhasca & Byrne 1994). However, it is less well under-
stood whether these effects are caused by biodiversity
per se
or by the presence of specific plant types
(Bengtsson 1998), but the degree of specificity prob-
ably depends on the mechanism involved. For
instance, the reduced herbivory hypothesis, also called
associational refuges (Pfister & Hay 1988) or associ-
ational resistance (Tahvanainen & Root 1972), may
be caused by general plant characteristics such leaf
shape or colour (Brown & Lawton 1991), or by species-
specific characteristics such as unique plant odour
compounds. The importance of different mechanisms
has rarely been evaluated, but associational refuges
appear to arise usually because of non-host plant
interference with the herbivore’s host-finding process
(Andow 1991; Finch & Collier 2000; Risch, Andow &
Altieri 1983).
In this paper we examine the species specificity of
one particular associational refuge, and the underlying
mechanisms. We examine the potential of olfactory
repellence (a species-specific mechanism) and visual
masking (a general mechanism) in a system with the
forb
Lythrum salicaria
and three monophagous her-
bivore species. The background is an earlier study
showing that
L. salicaria
experiences reduced her-
bivory, and an increased seed set by over a magni-
tude, when associated with the aromatic shrub
Myrica
gale
(Hambäck, Ågren & Ericson 2000). That study
also showed that reduced herbivory was probably
caused by host-finding interference, where the
presence of
Myrica
affected the host-finding process
by the monophagous leaf-feeding beetle
Galerucella
calmariensis
but not by the flower-feeding weevil
Nanophyes marmoratus
. In that study we were unable
to differentiate between an olfactory and a visual
interference mechanism. Here we tested the potential
of olfactory interference in laboratory trials with olfacto-
meters, and the specificity of the associational refuge
through a set of field experiments. In the laboratory,
†Author to whom correspondence should be addressed.
E-mail: peter.hamback@evp.slu.se

88
P. A. Hambäck,
J. Pettersson &
L. Ericson
© 2003 British
Ecological Society,
Functional Ecology
,
17
, 87–93
we examined (1) if olfactory cues are involved in
host-finding by the three main insect herbivores (
G
.
calmariensis
,
G
.
pusilla
and
N
.
marmoratus
) attacking
L
.
salicaria
; and (2) if the odour produced by
M. gale
either repels or masks the presence of
L
.
salicaria
. In
the field, we tested if the reduced herbivore abund-
ances and reduced herbivory on
L
.
salicaria
also
occurred when plants were placed in thickets of other
shrub species or in artificial thickets. In these experi-
ments we selected shrub species, and constructed arti-
ficial thickets, that were structurally similar to
Myrica
thickets. Hence any differences in the herbivore
response to these treatments are more likely to be a
consequence of olfactory rather than visual interfer-
ence of the host-finding process.
Materials and methods
 
The field work was performed in two regions. Two
experiments were performed in areas close to Umeå in
northern Sweden, Brännölandet (63
°
42
′
N, 20
°
25
′
E)
and Vitskärsudden (63
°
39
′
N, 20
°
17
′
E), and two
experiments were performed at the coast of Uppland
in southern Sweden, Jungfruholm (60
°
26
′
N, 18
°
15
′
E)
and Själgrundet (60
°
38
′
N, 17
°
45
′
E). All areas are
similar in structure, with shallow and rocky shorelines,
but have different mixtures of grasses, shrubs and
forbs. The areas are also characterized by land uplift,
and by ice-scouring in winter. In all areas,
L. salicaria
occurs from the lower part of the shore, close to the
average water table, to the
Alnus
fringe (
A
.
incana
in
the north and
A
.
glutinosa
in the south) in front of the
Spruce/Pine forest. All four shrub species included in
the study (
A
.
glutinosa
,
Hippophaë rhamnoides
,
M
.
gale
and
Salix myrsinifolia-phylicifolia
agg.) occur on the
upper part of the shore. However, different areas are
dominated by different shrub species:
S. myrsinifolia-
phylicifolia
agg. and
M. gale
grow at Brännölandet and
Vitskärsudden, while
A. glutinosa
and
H. rhamnoides
grow at Jungfruholm and Själgrundet.
 
Purple Loosestrife,
Lythrum salicaria
L. (Lythraceae),
is a perennial, insect-pollinated forb that is native to
Eurasia and has been introduced to North America
(Hultén & Fries 1986). In northern and central Sweden
it is mainly found on the shores of the Gulf of Bothnia.
Reproducing plants are 50 cm tall on average, and
produce several flowering shoots. Flower buds develop
in leaf nodes in the upper part of the flowering shoot.
In Fennoscandia,
L
.
salicaria
flowers for 6–8 weeks in
July–August and seeds mature 6–8 weeks after flower-
ing. Plant establishment is predominantly from seed.
Galerucella calmariensis
L. and
G
.
pusilla
(Coleoptera:
Chrysomelidae) are the main leaf-feeding herbivores
on Purple Loosestrife in study areas, and both species
are monophagous herbivores feeding on leaves and
flower buds. In the southern area (Uppland) both
species occur, although
G
.
pusilla
is the more abundant
species, while
G
.
calmariensis
is the only species
occurring in the northern area. The two species have
a similar life history, the only apparent difference
being that
G
.
calmariensis
adults are slightly larger.
Their life cycle is univoltine and the adults overwinter
in the litter (Hight
et al
. 1995). In southern Sweden
adults appear in mid-May; in northern Sweden they
appear in mid-June. The adults feed on leaves and lay
their eggs in batches on the stem and on the lower leaf
surface. The larvae hatch 7–10 days later, feed on
leaves and flower buds for 2–3 weeks, then pupate in
the ground. Feeding can cause extensive damage to the
host plant (Blossey 1992; Hambäck
et al
. 2000), and
both species are presently used as biocontrol agents for
Purple Loosestrife in North America (Hight
et al
.
1995). There is no distinct dispersal period for either
species, but dispersal could occur throughout the
growing season.
Nanophyes marmoratus
Goeze (Coleoptera: Curcu-
lionidae) is the main flower predator on Purple Loos-
estrife in the study area. It feeds solely on
L
.
salicaria
flowers and has a univoltine life cycle (Blossey &
Schroeder 1995). Adults appear by the end of June,
and the newly emerged adults feed on young leaves and
flower buds. The females oviposit on young flower
buds. The emerging larva consumes the reproductive
parts of the flower and pupates at the bottom of the
bud (Blossey & Schroeder 1995). While more than one
egg may be deposited in each flower, it is unusual for
more than one larva to reach pupation.
 
The abilities of
G
.
calmariensis
,
G
.
pusilla
and
N
.
mar-
moratus
to locate host plants through olfactory cues,
and their respective responses to
M
.
gale
, were exam-
ined using two-armed olfactometers (Fig. 1; for a
detailed description of the methodology see Petterson
et al.
1998). The olfactometers were made of perspex
Fig. 1. Two-armed olfactometer used to investigate insect
responses to plant odours. The arena for the beetle consists of
areas labelled A–C, and plants were placed in a chamber at
one or both ends (labelled empty vs plant). To create a
unidirectional odour flow, the centre of the olfactometer was
connected to an air-suction apparatus. At the start of a trial,
beetles were released in the arena centre (B) and positions
within the arena were then recorded every 3 min. To prevent
beetles reaching plants, a fine net covered the tube connecting
the behavioural arena with plant chambers.

89
Associational
refuges and
Galerucella
© 2003 British
Ecological Society,
Functional Ecology,
17, 87–93
and consisted of three layers with an arena cut out in
the middle sheet. This arena had a four-sided central
zone (2·5 × 2·5 cm) and two tapered arm zones, each
3·8 cm long and ending with a 0·6 cm broad space
where odour sources were connected with Teflon
tubing via 4 mm holes through the periphery. An air
stream over the arena was created by attaching a water
pump to the centre of the arena, and the air flow was
regulated to approximately 3 ml s−1. This was sufficient
to give an even cover of the arm zone with the applied
odour.
Each trial was run as follows. At the start of the trial,
one beetle was placed in each olfactometer and was
allowed to move for 5 min before measurements
started. Following this acclimatization period, we
recorded the position of the beetle every third minute
for 60 min (20 recordings). At each recording we
scored one of three positions, where the beetle was in
either arm (Fig. 1, areas A and C) or in the centre (area
B). If an individual was immobile (moved less than
5 mm) for more than three consecutive periods, this
individual was excluded and we tested a new indi-
vidual. For each stimulus we tested 36 individuals of G.
pusilla (18 males and 18 females) and 18 individuals of
G. calmariensis and N. marmoratus (females only), and
the same individual was used only once for each
stimulus. We found no sex differences, and therefore
present only pooled data. To test beetle responses
we used a Wilcoxon sign-rank test. The replicate in
this test was the individual beetle, and the test was
performed on the number of intervals that each
individual was recorded in either arm. All recordings of
an individual in the central part were excluded from
the analysis. Olfactometers were cleaned between runs,
first with water and mild detergent, then with 70% ethanol.
To test the hypothesis that olfactory interference
was involved in reduced herbivory on L. salicaria when
plants grew in Myrica thickets, we used the follow-
ing combinations of stimuli: (i) Lythrum–control;
(ii) Myrica–control; (iii) Lythrum/Myrica–control;
(iv) Lythrum/Myrica–Lythrum. We also included a zero
treatment, control–control, to examine the possibility
of unexpected responses caused by other aspects of the
laboratory environment, and a humidity control, wet
vs dry filter paper, to examine the possibility of attrac-
tion to water vapour emitted by plants. The possible
outcomes of the olfactory experiment are as follows. If
olfactory cues are involved in the host-finding process,
we would observe attraction to the Lythrum side
in experiment (i). If Myrica is repellent (active
avoidance), we would observe avoidance of the Myrica
side in experiments (ii) and (iv). If Myrica is masking
the presence of Lythrum (confusion), we would
observe zero response in experiment (ii), attraction to
the Lythrum side in experiment (iv), and reduced
attraction in experiment (iii) compared with (i). As these
hypotheses are nested, we performed a trial only if
previous trials indicated that this was necessary to
distinguish hypotheses.
 
We examined whether reduced herbivory on Purple
Loosestrife occurred only in thickets of M. gale, or if
the same response occurred when plants grew in thickets
of other shrub species that were unrelated to M. gale
but had a similar physical structure. We placed paired
transplants of L. salicaria in and outside similarly
sized thickets (0·5–5 m2) of Salix (Vitskärsudden in the
northern area, initiated 25 June 2001, N = 29); A. glu-
tinosa (Själgrundet in the southern area, initiated 4
June 2001, N = 20); and H. rhamnoides (Jungfruholm
in the southern area, initiated 2 June 2001, N = 28). We
also placed paired transplants in and outside artificial
thickets (Jungfruholm, initiated 2 June 2001, N = 28;
Brännölandet, initiated 25 June 2001, N = 28). Artifi-
cial thickets (0·5 m2) were constructed from a coarse
plastic net with tied-on green cotton stripes (2 cm
wide) up to a height similar to that of natural shrubs
(30–50 cm). If differences in herbivore abundance
between potted L. salicaria in and outside thickets are
equal for thickets of all four shrub species and for
artificial thickets, this indicates that the associational
refuge on L. salicaria is general. If olfactory inter-
ference of the host-finding process is the mechanism
underlying reduced herbivory on L. salicaria in
Myrica thickets, we would expect this difference to
disappear when using artificial thickets.
The experimental plants originated either from a
population 15 km north of Brännölandet (for experi-
ments in the northern region), or from a population
20 km north-east of Jungfruholm (for experiments in
the southern region). Plants had been grown from
seeds in experimental gardens at either Umeå or Uppsala,
and were 2–4 years old at the time of the experiment.
To prevent possible negative effects due to nutrient
competition from the shrub, or positive effects due to
nitrogen release from N-fixing nodules on the roots of
Myrica, Hippophaë or Alnus, plants were transplanted
in pots. The potted plants were paired based on size
at the time of transplantation to the field site. One
plant in each pair was placed in a natural or artificial
thicket, while the second plant was placed 2 m away
and outside the thicket. All potted plants were watered
regularly throughout the experiment.
On all potted plants we counted the number of G.
calmariensis and G. pusilla (adult beetles and eggs)
once, twice or three times, all within 3 weeks of trans-
plantation. We did not differentiate between the two
Galerucella species, as this is difficult for adults and
eggs in the field. In two areas (Jungfruholm and
Brännölandet) we also counted the number of adult N.
marmoratus. Finally, at Jungfruholm we counted the
number of Nanophyes larvae and pupae and the
number of flowers. Larvae and pupae of Nanophyes
were counted by collecting and dissecting all flowers
on each plant. This procedure is laborious and has
to be performed on fresh flowers. It was therefore
performed only in one area.

90
P. A. Hambäck,
J. Pettersson &
L. Ericson
© 2003 British
Ecological Society,
Functional Ecology,
17, 87–93
The differences in the number of Galerucella and
Nanophyes between potted L. salicaria in and outside
thickets was tested by the Wilcoxon sign-rank test. Prior
to testing we excluded double-zeros – where num-
bers were zero in both treatment and control within a
pair. The respective tests were performed on numbers
after 5 days (Galerucella adult); after about 2 weeks
(Galerucella eggs); after about 4 weeks (Nanophyes
adults); and after 2 months (Nanophyes larvae/pupae).
These periods depended on the species’ life histories.
Results
 
The zero treatment indicated no significant difference
between the two arms of the olfactometer (z = 0·04,
N = 18, P > 0·1): no artefact was inadvertently
included in the olfactometer experiment, and there
was no response to humidity for either Galerucella spe-
cies (Gp: z = 0·28, Gc: z = 0·45, N = 17, P > 0·1 for
both species). However, we found a negative response
to humidity for N. marmoratus (z = 2·0, N = 19,
P < 0·05). In the olfactometer treatments with plants,
the two Galerucella species showed similar responses
(Fig. 2). Neither species responded to the odour of
their host plant L. salicaria (Gp: z = 0·87, N = 33,
P > 0·1; Gc: z = 0·08, N = 18, P > 0·1), and both
species were attracted to the odour of M. gale (Gp:
z = 3·54, N = 38, P < 0·001; Gc: z = 2·20, N = 18,
P < 0·05). This indicates that olfactory responses were
less important in the initial host-finding process of G.
calmariensis and G. pusilla, and therefore that olfactory
interference by M. gale is less likely. As a consequence,
we did not perform experiments (iii) and (iv), which
were aimed only at separating two olfactory interference
mechanisms, masking and repellence. The response
of N. marmoratus was different from that of Galerucella.
The weevil responded positively to odour from Lythrum
(z = 2·28, N = 36, P < 0·05) and not to that of Myrica
(z = 0·20, N = 18, P > 0·1).
 
The response by Galerucella was similar for all treat-
ments (Fig. 3), although tests were hard to perform for
adult numbers because of many double zeroes. For
adults, numbers were reduced in artificial thickets
in the northern area (Brännölandet, z = 2·91, N =
21, P < 0·01), but a similar trend was not significant
for artificial thickets (z = 1·06, N = 8, P > 0·1), and
was marginally significant for Hippophaë thickets
(z = 1·63, N = 9, P = 0·1) in the southern area (Jung-
fruholm). At other sites, numbers were too few for a
meaningful test. For eggs, numbers were reduced in
Hippophaë thickets (z = 3·33, N = 18, P < 0·001);
Alnus thickets (z = 2·20, N = 7, P < 0·05); Salix thickets
(z = 4·27, N = 25, P < 0·001); and artificial thickets in
both the northern (Brännölandet, z = 3·20, N = 26,
P < 0·001) and southern area (Jungfruholm, z = 2·41,
N = 20, P < 0·05). The response by N. marmoratus was
different from the Galerucella response (Fig. 4), but
again tests were hard to perform for adult numbers.
For adults, numbers were higher in both Hippophaë
(z = 2·10, N = 6, P < 0·05) and artificial thickets
Fig. 2. Responses of Galerucella pusilla (Gp), G. calmariensis
(Gc) and Nanophyes marmoratus (Nm) to the odour of the
host plant Lythrum salicaria and to the non-host Myrica gale
(***, P < 0·001; *, P < 0·05). Trials were performed in a two-
armed olfactometer (Fig. 1).
F
ig. 3. The number of Galerucella calmariensis/G. pusilla [adults (a); eggs (b),
m
ean ± SE; ***, P < 0·001; **, P < 0·01; *, P < 0·05] on the host plant Lythrum
s
alicaria when host plants were placed in either an open control area (open bar) or
inside a thicket [solid bar; thickets include Myrica gale, Alnus glutinosa, Hippophaë
r
hamnoides, Salix myrsinifolia-phylicifolia agg., and artificial thickets at Brännölandet
(
1) and Jungfruholm (2)]. Data for the M. gale study are from Hambäck et al. (2000).

91
Associational
refuges and
Galerucella
© 2003 British
Ecological Society,
Functional Ecology,
17, 87–93
(z = 2·10, N = 6, P < 0·05) in the southern area, while
no difference was observed between treatments in the
artificial thicket experiment in the northern area
(Brännölandet, z = 0·10, N = 6, P > 0·1). The experi-
ment at Brännölandet ended prematurely due to
destruction by waves, and adults were therefore scored
prior to flowering. Larval and pupal numbers of Nan-
ophyes showed a pattern similar to adult numbers,
where numbers were higher in Hippophaë thickets
(z = 3·78, N = 19, P < 0·001) and in artificial thickets
(z = 1·99, N = 14, P < 0·05) at Jungfruholm. Finally,
flower numbers were higher in both Hippophaë thickets
(z = 2·62, N = 21, P < 0·01) and artificial thickets (z =
2·50, N = 18, P < 0·05) at Jungfruholm, and the effect
size was similar to that for Nanophyes larvae/pupae.
Discussion
This study shows that the associational refuge for L.
salicaria is not specific to thickets of M. gale, but a
similar refuge is provided by three other shrub species.
The reduced herbivore abundance of the two leaf-
feeding beetles G. calmariensis and G. pusilla was
quantitatively similar for thickets with M. gale, A. glu-
tinosa, H. rhamnoides, S. myrsinifolia-phylicifolia agg.,
and artificial thickets (Fig. 3). It is therefore also likely,
even though this was measured only in three of six
experiments, that the reduced herbivory translates into
an increased reproductive output by L. salicaria. The
probable reason for this generalistic associational ref-
uge is that the two Galerucella species do not appear to
use olfactory cues for locating undamaged Lythrum
individuals, and are not repelled by odours emitted
from M. gale (Fig. 2). It appears that the associational
refuge for L. salicaria inside shrubby thickets is due to
visual interference, an interference mechanism that is
equally well provided by all shrub species investigated
and by the artificial shrubs.
The fact that olfactory cues are not used for primary
host plant selection by G. calmariensis and G. pusilla
suggests that undamaged hosts are located by Galeru-
cella individuals mainly by a more-or-less random
movement (Grevstad & Herzig 1997). Once a host
plant has been located, and verified through gustatory
stimuli, egg-laying commences. The presence of non-
host plants increases the movement rates and therefore
also increases the probability of moving away from the
host plant (similar to the reasoning of Kareiva & Odell
1987; Finch & Collier 2000). The resulting asymmetrical
distribution of Galerucella individuals in and outside
shrubs may then be reinforced through aggregation,
occurring as a consequence of beetle attraction to
damaged Lythrum individuals (P.A.H., J.P. and S. Al
Abassi, unpublished results; see also Grevstad & Herzig
1997; Peng & Weiss 1992).
The response by Galerucella individuals to non-host
plant neighbours contrasts with that of N. marmoratus.
Nanophyes showed dramatic increases in density inside
shrubs (Fig. 4), and appeared to use olfactory cues to
a greater extent than Galerucella individuals (Fig. 2).
It is probable that the presence of non-host plant
neighbours therefore has no effect on host-finding by
Nanophyes, because they are better able than Galeru-
cella to differentiate between undamaged host and
non-host plants during the host-finding process. This
lack of response would explain a zero difference on
plants in and outside shrubby thickets, while the higher
Nanophyes density inside thickets demands an addi-
tional explanation. The most likely mechanism is that
the higher flowering frequency of L. salicaria in the
thickets (Fig. 4), an effect caused by the reduced
abundance of and herbivory by Galerucella adults
and larvae (Hambäck et al. 2000), is a strong attractant
to Nanophyes individuals.
These mechanisms suggest that the development of
a general theory on associational refuges which would
enable us to predict when plant neighbours reduce her-
bivore numbers should take account of the herbivore
host-finding process. This study suggests that reduced
herbivory in the presence of non-host plant neigh-
bours is mainly achieved for herbivore species that rely
on visual or gustatory cues in the host-finding process.
Fig. 4. Number of Nanophyes marmoratus [adults (a); larvae/pupae (b), mean ± SE,
***, P < 0·001; **, P < 0·01; *, P < 0·05] on the host plant Lythrum salicaria, and
flower number (b) when host plants were placed in an open control area (open bar) or
inside a thicket [solid bar; thickets include Myrica gale, Hippophaë rhamnoides and
artificial thickets at Brännölandet (1) and Jungfruholm (2)]. Data for the M. gale study
are from Hambäck et al. (2000).

92
P. A. Hambäck,
J. Pettersson &
L. Ericson
© 2003 British
Ecological Society,
Functional Ecology,
17, 87–93
However, the generality of this hypothesis is difficult to
evaluate, as few previous studies have combined field
measurements of herbivory rates in the presence and
absence of non-host plant neighbours with obser-
vations of underlying behavioural mechanisms. An
exception is Finch & Collier (2000), who found
reduced herbivory by cabbage root flies (Delia radi-
cum) in mixed croppings. By observing the egg-laying
behaviour of individual root flies, they argued that egg-
laying was initiated only following a repeated number
of contacts with host plants, and that the presence of
non-host plants disturbed this process leading to rejec-
tion of an egg-laying site. Hence, a larger non-host
plant density near cabbage plants would reduce egg-
laying by root flies. This response contrasts with that
of diamondback moth (Plutella xylostella), which
approaches host plants differently, and in this case
intercropping is a less effective means of reducing
attack rates on cabbage. It would be of interest to
examine the host-finding ability and the mechanisms
involved for a larger set of herbivore species, and to
relate these to successes and failures in intercropping
experiments. It is probable that this knowledge would
aid in understanding the large variability observed in
these studies (Andow 1991) and assist in the develop-
ment of more effective intercropping systems.
In a wider context, if generalistic associational
refuges are more common than species-specific associ-
ational refuges, as suggested in this study, there will be
consequences for the interpretation of recent studies
relating plant diversity and productivity. Generalistic
associational refuges through physical interference
would cause herbivory to level off quickly with increas-
ing plant diversity. Once a sufficiently large non-host
plant neighbour density is established, it is unlikely
that additional plant neighbours cause further reduc-
tions in herbivory. Hence, a productivity increase
through reduced herbivory is also likely to level off at
moderate plant diversities. At present few data are
available to support this conclusion, but future studies
that mix herbivore exclusions with plant diversity
manipulations may provide important insights. It
would be helpful if future studies on plant diversity
effects on herbivore abundance and herbivory were
accompanied by mechanistic studies. As shown here,
different herbivore species may respond negatively and
positively to the same treatment, and this variability
may, at least partly, be understood by examining
behavioural responses by the different species and the
consequent feedbacks through responses by other
species within the community.
Acknowledgements
This paper was improved by comments from Jon
Ågren, Richard Hopkins and an anonymous reviewer.
We thank Markus Brage who assisted in both field and
laboratory studies. Funding was provided by Stiftelsen
Oskar och Lilli Lamms Minne.
References
Andow, D.A. (1991) Vegetational diversity and arthropod popu-
lation response. Annual Review of Entomology 36, 561–586.
Bengtsson, J. (1998) Which species? What kind of diversity?
Which ecosystem function? Some problems in studies of
relations between biodiversity and ecosystem function.
Applied Soil Ecology 10, 191–199.
Blossey, B. (1992) Impact of Galerucella pusilla and G. cal-
mariensis (Coleoptera: Chrysomelidae) on field populations
of purple loosestrife ( Lythrum salicaria ). Proceedings of the
Eighth International Symposium on Biological Control of
Weeds (eds E.S. Delfosse. & R.R. Scott), pp. 27–31. DSIR/
CSIRO, Melbourne.
Blossey, B. & Schroeder, D. (1995) Host specificity of three
potential biological weed control agents attacking flowers
and seeds of Lythrum salicaria (Purple loosestrife). Biolo-
gical Control 5, 47–53.
Brown, V.K. & Lawton, J.H. (1991) Herbivory and the evolu-
tion of leaf size and shape. Philosophical Transactions of the
Royal Society of London Series B 333, 265–272.
Finch, S. & Collier, R.H. (2000) Host-plant selection by
insects – a theory based on ‘appropriate/inappropriate
landings’ by pest insects of cruciferous plants. Entomologia
Experimentalis et Applicata 96, 91–102.
Grevstad, F.S. & Herzig, A.L. (1997) Quantifying the effecs
of distance and conspecifics on colonization: experiments
and models using the loosestrife leaf beetle, Galerucella cal-
mariensis. Oecologia 110, 60 –68.
Hambäck, P.A., Ågren, J. & Ericson, L. (2000) Associ-
ational resistance: insect damage to purple loosestrife
reduced in thickets of sweet gale. Ecology 81, 1784–1794.
Hector, A., Schmid, B., Beierkuhnlein, C., Caldeira, M.C.,
Diemer, M., Dimitrakopoulos, P.G., Finn, J.A., Freitas,
H., Giller, P.S., Good, J., Harris, R., Högberg, P., Huss-
Danell, K., Joshi, J., Jumpponen, A., Körner, C., Leadley,
P.W., Loreau, M., Minns, A., Mulder, C.P.H., O’Donovan,
G., Otway, S.J., Pereira, J.S., Prinz, A., Read, D.J., Scherer-
Lorenzen, M., Schulze, E.D., Siamantziouras, A.S.D.,
Spehn, E.M., Terry, A.C., Troumbis, A.Y., Woodward,
F.I., Yachi, S. & Lawton, J.H. (1999) Plant diversity and
productivity experiments in European grasslands. Science
286, 1123–1127.
Hight, S.D., Blossey, B., Laing, J. & Declerck-Floate, R.
(1995) Establishment of insect biological control agents
from Europe against Lythrum salicaria in North America.
Environmental Entomology 24, 967–977.
Hultén, E. & Fries, M. (1986) Atlas of North European Vas-
cular Plants North of the Tropic of Cancer, I–III. Koeltz,
Königstein.
Kareiva, P. & Odell, G. (1987) Swarms of predators exhibit
‘preytaxis’ if individual predators use area-restricted
search. American Naturalist 130, 233–270.
Koricheva, J., Mulder, C.P.H., Schmid, B., Joshi, J. & Huss-
Danell, K. (2000) Numerical responses of different trophic
groups of invertebrates to manipulations of plant diversity
in grasslands. Oecologia 125, 271–282.
Mulder, C.P.H., Koricheva, J., Huss-Danell, K., Högberg, P.
& Joshi, J. (1999) Insects affect relationships between plant
species richness and ecosystem processes. Ecology Letters
2, 237–246.
Peng, C. & Weiss, M.J. (1992) Evidence of an aggregation
pheromone in the flea beetle, Phyllotreta cruciferae (Goeze)
(Coleoptera: Chrysomelidae). Journal of Chemical Ecology
18, 875–884.
Petterson, J., Karunaratne, S., Ahmed, E. & Kumar, V. (1998)
The cowpea aphid, Aphis craccivora, host plant odours and
pheromones. Entomologia Experimentalis et Applicata 88,
177–184.
Pfister, C.A. & Hay, M.E. (1988) Associational plant
refuges: convergent patterns in marine and terrestrial

93
Associational
refuges and
Galerucella
© 2003 British
Ecological Society,
Functional Ecology,
17, 87–93
communities result from differing mechanisms. Oecologia
77, 118–129.
Risch, S.J., Andow, D. & Altieri, M.A. (1983) Agroecosystem
diversity and pest control: data, tentative conclusions, and new
research directions. Environmental Entomology 12, 625 – 629.
Tahvanainen, J.O. & Root, R.B. (1972) The influence of vegeta-
tional diversity on the population ecology of a specialized
herbivore, Phyllotreta cruciferae (Coleoptera: Chrysomelidae).
Oecologia 10, 321–346.
Tilman, D., Wedin, D. & Knops, J. (1996) Productivity and
sustainability influenced by biodiversity in grassland eco-
systems. Nature 379, 718–720.
Tonhasca, A. & Byrne, D.N. (1994) The effects of crop diver-
sification on herbivorous insects: a meta-analysis
approach. Ecological Entomology 19, 239–244.
Received 5 September 2002; revised 13 September 2002;
accepted 20 September 2002