12MachialAFE14.pdf

Page 1 of 9

Agricultural and Forest Entomology (2012), 14, 286–294 DOI: 10.1111/j.1461-9563.2012.00568.x

The role of vision in the host orientation behaviour of Hylobius

warreni

Laura A. Machial, B. Staffan Lindgren and Brian H. Aukema∗

Natural Resources and Environmental Studies Institute, University of Northern British Columbia, 3333 University Way, Prince George, British

Columbia V2N 4Z9, Canada, and ∗Department of Entomology, University of Minnesota, 217 Hodson Hall, 1980 Folwell Avenue, St Paul, MN

55118, U.S.A.

Abstract 1 Visual stimuli, often in combination with olfactory stimuli, are frequently important

components of host selection by forest-dwelling phytophagous insects.

2 Warren root collar weevil Hylobius warreni Wood (Coleoptera: Curculionidae) is a

native insect in western Canada, where larvae feed primarily on lodgepole pine Pinus

contorta and can girdle and kill young trees. This weevil is an emerging problem in

areas heavily impacted by mountain pine beetle Dendroctonus ponderosae Hopkins.

3 No olfactory attractants have been identified for this insect, making monitoring and

management difficult. Thus, we investigated the role of vision in the host-finding

behaviour of Warren root collar weevil in the absence of known olfactory cues.

4 We conducted three experiments in field enclosure plots aiming to characterize

aspects of host-finding behaviour by adult Warren root collar weevil.

5 We found that both male and female weevils were readily attracted to vertical plastic

silhouettes in the shape of a trunk, crown or tree at distances of less than 4 m. This

pattern of attraction persisted over 2 years in two slightly different study designs.

Blinding the insects removed their ability to orient to these silhouettes, indicating

that host-finding behaviour has a strong visual component. The use of different

colour trunks and crowns (black, white and green) did not change the patterns of

attraction of the insects to the silhouettes.

6 Exploiting visual attraction in this walking insect may present a new management

tool in forest protection strategies.

Keywords Host selection, host-finding behaviour, silhouettes, vision, Warren root

collar weevil.

Introduction

Host selection by phytophagous insects is characterized by

two activities: host location and host assessment (Dethier,

1983). Locating hosts requires finding appropriate habitat and

then identifying a host plant, potentially among other nonhost

vegetation (Dethier, 1983; Huber et al., 2000; Raffa, 2001;

Bernays, 2003). Insects utilize various strategies to maximize

their chances of encountering an appropriate plant, such as

increasing activity, moving randomly, turning frequently and/or

responding to various host stimuli with an orderly sequence of

behaviour patterns (Dethier, 1983). Olfactory and visual stimuli

are both perceived as attractive cues by many insects when

Correspondence: Laura Machial. Tel.: (651) 408-4239;

fax: (612) 625-5299; e-mail: lauramachial@gmail.com

initially locating plants on which to feed and oviposit (Bernays

& Chapman, 1994; Bernays, 2003).

Although both types of cues have been shown to be

integrated by phytophagous insects, including many types of

forest-dwelling arthropods (Borden et al., 1985; Strom et al.,

1999, 2001; Bjorklund ̈ et al., 2005; Campbell & Borden, 2006a,

b), olfactory cues are frequently viewed as the more important.

Not unexpectedly, the role of olfaction has received the most

attention in studies of how insects select their hosts. The use of

visual stimuli in detecting plants has been reported for several

species (Prokopy & Owens, 1983; Reeves, 2011), although

few insects were found to use visual cues in the absence of

olfactory cues (Stenberg & Ericson, 2007; Reeves et al., 2009;

Reeves, 2011).

Warren root collar weevil Hylobius warreni Wood

(Coleoptera: Curculionidae) is a phytophagous insect that is

Page 2 of 9

Host-finding behaviour of H. warreni 287

native throughout Canada’s boreal forests (Cerezke, 1994).

Adult weevils can reach 5 years of age. The weevils cannot fly,

and thus they encounter hosts as they walk along the forest floor

(Grant, 1966; Cerezke, 1994). In British Columbia, Canada, the

weevil’s primary host is lodgepole pine Pinus contorta var. lat- ifolia (Warren, 1956; Wood, 1957; Grant, 1966). Adults ascend

conifer trees at dusk to feed on the branches, bark and needles.

Feeding by adult Warren root collar weevil typically does not

cause tree mortality (Warren, 1956; Warner, 1966). Larval feed- ing, however, may cut into the cambial and xylem tissues of the

host’s large lateral roots or root collar (Warren, 1956; Cerezke,

1994). In young trees, complete girdling can be accomplished

by as few as one to three larvae. In trees with larger diameters,

many more weevils are required to complete girdling, although

insects rarely occur at such high densities. Warren root collar

weevils can cause mortality in stands as old as 30 years of age,

although peak mortality occurs in stands aged 5–10 years of

age (Cerezke, 1994).

The current epidemic of mountain pine beetle Dendroctonus

ponderosae Hopkins has killed over 630 million m3 of mature

lodgepole pine in British Columbia and Alberta since 1998 and

is predicted to kill approximately 67% of mature pine in British

Columbia by 2016 (Aukema et al., 2006; Walton, 2010). As a

result, salvage logging and reforestation has begun to shift the

age structure and species composition of large areas of forest.

Recently, studies have reported increased mortality to young

trees caused by Warren root collar weevil (Robert & Lindgren,

2006), especially in areas where harvested and reforested lands

are located adjacent to stands of mature lodgepole pine heavily

affected by mountain pine beetle where the timber has not

been removed. Such patterns are consistent with the hypothesis

that weevils are migrating from forests with depleted host

pools to replanted areas in search of new hosts (Klingenberg

et al., 2010b).

Despite the increased importance of the insect, little is

known about the host selection mechanisms of Warren root

collar weevil. In closely-related species of Hylobius, volatiles

of host terpenes and ethanol play prominent roles in host

attraction (Tilles et al., 1986a, b; Raffa & Hunt, 1988).

Attraction to such compounds has been exploited to increase

trapping efficiency and has greatly aided the development

of management plans for pine weevils (Nordlander, 1987;

Raffa & Hunt, 1988; Hunt & Raffa, 1989; Rieske & Raffa,

1993). Studies aiming to determine whether Warren root

collar weevils are similarly attracted to host volatiles have

yielded inconclusive results (K. Sambaraju and B. S. Lindgren,

personal communication; Duke & Lindgren, 2006). In these

field and laboratory olfactometer studies, weevils failed to

show responses to chemical stimuli such as monoterpene

components that characterize lodgepole pine. Moreover, no

clear links have been found between tree chemistry and

susceptibility to attack by Warren root collar weevil (Duke &

Lindgren, 2006).

It is possible that vision plays a prominent role in host

orientation for Warren root collar weevil. In a review on the

ecology, behaviour and management of the weevil, Cerezke

(1994) commented that host selection was probably visual

in nature and thus influenced by silhouettes. This hypothesis

was proposed after noticing that the capture frequency of

adult weevils was positively correlated with an increasing tree

diameter (Cerezke, 1994). Moreover, a number of studies have

observed that weevils orient and move toward two-dimensional

silhouettes in laboratory settings (Hoover, 2000; Horning &

Lindgren, 2002). These findings are consistent with those found

for other conifer-feeding insects, including other pine weevil

species that integrate visual cues in the presence of olfactory

cues. For example, the pales weevil H. pales Herbst is more

attracted to traps baited with ethanol and turpentine that have

white silhouettes than to traps with black or green silhouettes

(Hunt & Raffa, 1991).

The present study aimed to provide additional information

concerning the role of vision in the host-finding behavior of

Warren root collar weevil. We investigated the hypothesis that

vision plays a role in initial steps of the weevils’ host finding

by exploring three questions. First, are weevils attracted to

visual cues? If so, does blinding the insects remove their host- finding ability? Finally, does colour affect locomotor response

to vertical silhouettes?

Materials and methods

Site set-up

During the summers of 2009 and 2010, we conducted

three experiments aiming to determine the role of vision in

host selection by Warren root collar weevil. In 2009, host- seeking experiments were conducted in 16 square, outdoor

bioassay plots measuring 1.5 × 1.5 m in an approximately

20 × 20 m area with bentgrass, Agrostis sp., vegetation in

Prince George, British Columbia, Canada (53◦

43.2N,

122◦

39.6W). Each plot was surrounded by a 1-m tall

polypropylene mesh wrapped around wooden corner stakes.

The bottom 15 cm of the mesh was lined using duct tape

with a slippery fluorocarbon polymer (2009: AD1070, AGC

Chemicals Americas, Inc., Bayonne, New Jersey; 2010: Teflon

PTFE, DuPont, Wilmington, Delaware) to prevent the weevils

from climbing the mesh and escaping from the plot (Bjorklund, ̈

2009). In addition, the bottom of the mesh was pinned to the

ground with nails (length 5 cm) as a further measure to prevent

weevil escape. The mesh and stakes were painted white to

maximize the contrast of the silhouette treatment, inside the

plot, against the background. Plots were located a minimum of

2 m apart.

Within each plot, two plastic flower pots (diameter 25 cm,

depth 10 cm) were installed as pitfall traps. The traps were

located 15 cm from the north and south sides of each plot such

that the distance between the centres of the traps was 95 cm

(Fig. 1A). Pitfall traps were placed so that their tops were flush

with the ground. The top 4 cm inside the pitfall traps was coated

with polymer to prevent captured weevils from climbing the

sides and escaping.

In 2010, the study site was located at the Prince George

Tree Improvement Station (53◦

18N, 122◦

4W). This

permitted enlargement of the plots to 2 × 4 m, spacing centres

of the two pitfall traps in each arena 1 m from the fencing

at both sides of the plot, and increasing the distance between

centres of pitfall traps within a plot to 2 m (Fig. 1B). We judged

this distance of 2 m to be more representative of the mean

Page 3 of 9

288 L. A. Machial et al.

(A) (B)

Figure 1 Schematic diagram of the design of outdoor bioassay plots

used to investigate the role of vision in initial steps of adult Warren root

collar weevils’ host finding. Each plot had a treatment trap and an empty

trap, except the control treatment, which had two empty traps. (A) Layout

of 16 bioassay plots used in 2009 in Prince George, British Columbia,

Canada (53◦

43.2N, 122◦

39.6W). (B) Layout of 20 bioassay

plots used in 2010 at the Prince George Tree Improvement Station,

Prince George, British Columbia, Canada (53◦

18N, 122◦

4W).

Five weevils were released per plot.

distance that Warren root collar weevils can move in a night

(Cerezke, 1994). We constructed 20 plots in 2010.

In both years, we tested weevil responses by erecting a

dummy host-tree, which provided a visual stimulus in the form

of a vertical silhouette in one of the two pitfall traps, chosen

randomly by a coin toss, within each arena. Different types of

silhouettes were used to emulate different potential hosts, and

thus defined the experimental treatment. The specific treatments

are described further below. The silhouette was erected in

the middle of the pitfall trap such that it did not contact the

ground outside of the trap, preventing the insects from climbing

it without first entering the pitfall trap. The treatments were

randomly assigned to the plots. Control treatments did not

receive a silhouette (i.e. both pitfall traps were empty in those

arenas).

Study organisms

Bjorklund funnel traps were made using tar paper (GAP ̈

Roofing Black #30 Roofing Felt, GAP Roofing, Inc., Pryor,

Oklahoma) and installed on 167 lodgepole pine trees in a

10–20-year-old lodgepole pine stand in 2009, and on 602 trees

in similar stands (7–25 years) in 2010 near Prince George,

British Columbia, Canada (53◦

N, 122◦

W) in accordance

with the protocol previously described by Bjorklund (2009). ̈

In 2009, there were two sampling periods, whereas, in 2010,

there was only a single, longer, sampling period. The traps

were checked each morning, and yielded 324 adult weevils on

11 May to 17 June and 13–28 July 2009, and 519 adult weevils

on 26 April to 18 August, 2010.

Captured weevils were retained in groups of eight in square

plastic containers (15 × 15 × 5 cm) with a piece of moist

paper towel and a few small lodgepole pine branches for food

(Toivonen & Viiri, 2006; Hopkins et al., 2009). Food was

changed twice per week, as required. The weevils were stored

in an environmental growth chamber under a LD 12 : 12 h

photocycle at 8 ◦

C and 75% relative humidity. The temperature

was set at 8 ◦

C to slow weevil metabolism (Toivonen &

Viiri, 2006).

The sex of the weevils was determined using two non- invasive techniques described previously by Ohrn ̈ et al. (2008).

Internal markings on the eighth sternite were also examined to

increase confidence in the accuracy of the identification (G. R.

Hopkins, M. D. Klingenberg and B. H. Aukema, unpublished

data). Male and female weevils were then divided into separate

plastic containers, again in groups of eight.

Experimental trials

In both years, all weevils utilized for an experiment had been

captured that season. We marked each insect because the

recapture rate of weevils released was not 100% and plots were

reused for numerous trials. Marking was conducted by etching

the elytra with a rotary drill as described by Klingenberg et al.

(2010a), comprising a technique that has been successfully

used on other ground-dwelling Coleoptera (Winder, 2004).

Etchings were then filled in with nontoxic latex-based paint

(Citadel Colour, U.K.) to accentuate markings (Klingenberg

et al., 2010a).

Before each experimental trial, weevils were placed in

clean plastic containers without food for a period of 24 h.

Containers were kept at ambient temperatures and exposed to

the natural photoperiod in the range 12.8–14.2 h/day in 2009

to 14.3–17.0 h/day in 2010 depending on the season (National

Research Council of Canada, 2011). For each experiment, five

weevils were released along the centre line; once per plot per

trial. This density reflects the highest weevil density commonly

seen on single hosts in 20–25-year stands (Cerezke, 1994).

Each trial ran 60 h. Pitfall traps were checked every 12 h for

the duration of the trial. When a weevil was found in a pitfall

trap, the replicate and treatment were recorded. The weevil

was subsequently removed from the experiment. Insects were

not reused.

Experiment 1: are weevils attracted to visual cues?

This experiment investigated the response of Warren root

collar weevils to four silhouette treatments: crown, trunk, tree

(crown + trunk) and control. The crown treatment consisted

of a plastic Christmas tree (height 138 cm, trunk diameter

3 cm). The trunk treatment consisted of a piece of acrylonitrile

butadiene styrene (ABS) pipe (height 90 cm, diameter 10 cm).

The tree treatment consisted of a piece of ABS pipe (height

90 cm, diameter 10 cm), with a plastic Christmas tree (height

138 cm) inserted to give the appearance of the pipe being

the trunk of the tree. The combined height of the ABS pipe

Page 4 of 9

Host-finding behaviour of H. warreni 289

and Christmas tree was approximately 188 cm as a result

of part of the Christmas tree being inserted into the ABS

pipe. The control treatment had no silhouettes in either pitfall

trap. In 2009, artificial Christmas trees were procured from

garage sales in the Prince George metropolitan area. In 2010,

the artificial trees were identical ‘Canadian pine’ specimens

(original height 200 cm, crown height 104.1 cm), which were

purchased from Wal-Mart (Canada). Despite the name, the trees

had a visual appearance of a unique hybrid of pine and spruce;

hosts acceptable to H. warreni. The trees had seven rows of six

to eight branches. The branches decreased in length from the

bottom of the tree to the top; the largest branches were 53 cm

and the shortest were 27 cm. The branches had 10–16 ‘shoots’,

and the shoots were covered in 3-cm long plastic ‘needles’. All

plastic trees were unboxed and exposed to the air for over 48 h

before use in experiments, although no associated odours were

detectable to the human nose during the experiment set-up.

Replicated trials were conducted from 25 to 28 August and

29 to 31 August 2009 and from 17 to 19 June and 23 to 25

June, 2010.

In 2009, a logistic mixed-effects analysis of variance was

used to examine the effect of silhouette treatments on the

number of weevils captured in control versus pitfall trap

treatments (a binary response). Plot was the unit of replication.

Fixed effects included terms for crown and trunk, as well

as their interaction, whereas a term for plot was modelled

as a random effect. Where significant differences between

treatments existed (using α = 0.05), means comparisons were

performed using protected t-tests (Carmer & Swanson, 1973).

In addition, a chi-square contingency analysis was conducted

to determine whether there was a difference in the number of

weevils caught in pitfall traps baited with the different treatment

types. In 2010, with expanded plot size, no weevils were caught

in empty traps within the treatment plots (see Results). A

lack of weevils in these traps created an analytical challenge

in estimating the proportion of weevils attracted to control

versus treatments and the standard errors of those estimates

in a binomial model framework because only treatments with a

silhouette elicited positive responses. Hence, in 2010, we used

chi-square contingency analysis to determine whether there

was a difference in the number of weevils caught in pitfall

traps with the different treatment types. All data analyses were

performed in r, version 2.12.2 (R Development Core Team,

2011).

Experiment 2: does blinding remove host-finding ability?

This experiment investigated whether vision was a mechanism

used by Warren root collar weevil in host orientation. Each

plot was set up with two pitfall traps as described for the

2010 experiments above. One pitfall trap contained an artificial

Christmas tree crown inserted into an ABS pipe to serve as

a tree silhouette (tree treatment); the other was left empty

(control). Five weevils were released in each of the 20 bioassay

plots as described previously. In half the plots, blinded weevils

were released; in the other half, nonblinded weevils were

released.

Weevils were blinded using nontoxic Elmer’s All Ceramic

and Glass Cement mixed with lamp black nontoxic acrylic paint

(Americana, Elmery’s, Columbus, Ohio). Blinding insects by

applying paint to their eyes has confirmed use of vision in

silhouette location by common field grasshoppers Chorthippus

brunneus Thunberg (Kral, 2008), as well as prey-finding

behaviour by a number of predaceous insects (Rossel, 1986;

Awan et al., 1989; Claver & Ambrose, 2001). The glue–paint

mixture was applied to the weevils’ eyes using a synthetic 000

S/H round paintbrush (Winsor & Newton University Series,

U.K). Trials were run from 12 to 14 September, 2009, 11 to

13 June, 2010 and 6 to 8 July, 2010.

In 2009, a likelihood ratio test (G-test) was used to examine

the effect of blinding on the number of weevils found in empty

pitfall traps versus pitfall traps underneath tree silhouettes. The

test statistic of the G-test is distributed according to a chi-square

distribution, and can be used when expected values are less than

five in the contingency table (Gotelli & Ellison, 2004).

As in Experment 1, the larger plots in 2010 resulted in very

few insects being captured in the empty controls (see Results).

Hence, we focused only on insects captured in pitfall traps with

a tree silhouette. When inspecting weevils post-capture under

a dissecting microscope, we found that some of the weevils’

eyes were only partially covered with paint, either as a result of

burrowing activity or scratching with tarsi. Because we could

not conclusively confirm that all weevils were 100% blinded,

we analyzed the likelihood of a weevil being captured in a

trap with a tree silhouette as a function of vision impaired by

paint versus nonblinded. In this logistic regression analysis, the

statistical test associated with estimate of the intercept divided

by its standard error yields a Z-statistic reflecting a test of

whether the probability of recovering a nonblinded weevil in a

silhouette baited trap was significantly different from 50% (i.e.

reflecting random movement) [the test of whether the intercept

was significantly different from zero, back-transformed by the

logit link function, exp0/(1 + exp0), reflects Ho: P(recovered

specimen has paint on its eyes) = 0.50]. Again, a term for

plot was included as a random effect. Data analysis was

performed in r, version 2.12.2 (R Development Core Team,

2011).

Experiment 3: are weevils attracted to colour?

Experiments looking at the effects of white/black trunks and

white/green crowns on Warren root collar weevil host selection

were conducted during the summer of 2010. Four treatments

were investigated. The treatments were ‘trees’ constructed with

a white trunk and a green crown, a white trunk and a white

crown, a black trunk and a white crown, and a black trunk and

a green crown. The white trunks were composed of polyvinyl

chloride (PVC) pipe (height 90 cm, diameter 10 cm), and black

trunks were composed of similar ABS pipe. The white crowns

were painted twice with white acrylic, latex, exterior, flat

paint (BEHR Premium Plus Ultra Pure White Co., Santa Ana,

California). The crowns of the green artificial Christmas trees

were left unpainted. Again, each treatment was characterized by

a silhouette placed into a pitfall trap opposing an empty pitfall

trap in the other side of the enclosure plot. Two replicates of

each of the four treatments were conducted at a time. Four trials

were conducted: 3–5 August, 9–11 August, 12–14 August and

21–23 August, 2010.

Page 5 of 9

290 L. A. Machial et al.

Before carrying out the present study, we conducted a

preliminary study investigating the effects of paint on weevil

behaviour. In that experiment, weevils were given a choice

between two similar tree silhouettes: one crown was painted

with a matching green colour of the original plastic branches,

whereas the other was left unpainted (= green). A logistic

regression model was used to determine whether paint affected

the number of weevils caught in the pitfall traps surrounding

both trees) (Z = −1.24, P = 0.215. Thus, we concluded that

paint, at least in the tree crowns, did not affect weevil behaviour

(L. Machial, unpublished data).

As a result of the tendency of weevils to avoid empty pitfall

traps in the larger plots in 2010, we again focused solely on

the number of weevils caught in pitfall traps surrounding the

four, coloured silhouettes. A chi-square contingency analysis

was conducted to determine whether there was a difference

among treatments. Data analysis was performed in r, version

2.12.2 (R Development Core Team, 2011).

Results

Experiment 1: are weevils attracted to visual cues?

Weevils demonstrated clear attraction to vertical silhouettes

in both 2009 and 2010, although the type of silhouette

eliciting attraction was not necessarily consistent between years

(Fig. 2). For example, in 2009, the proportion of weevils

falling into the two pitfall traps in the control plots was not

significantly different from 50% (i.e. a random pattern of

dispersion in the absence of vertical silhouettes). The crown

and trunk treatments did not differ from controls, although the

tree treatment that combined the trunk and crown silhouette

captured significantly more weevils than the associated empty

trap (F1,30 = 5.48; P = 0.0259; Fig. 2A). In addition, when

the proportion of weevils found in pitfall traps baited with

silhouettes was directly compared, more weevils were found

in pitfall traps baited with tree silhouettes than in crown,

trunk or the empty control pitfall trap (χ2 = 10, d.f. = 3,

P = 0.0186).

In 2010, proportionately more weevils were found in the

pitfall traps surrounding crown, trunk and tree silhouettes

than in the empty control pitfall traps (χ2 = 13.7, d.f. = 3,

P = 0.004; Fig. 2B). Indeed, not a single weevil was captured

in the empty trap in any of the plots containing a silhouette.

There was no preference for tree silhouettes over crown or

trunk silhouettes (χ2 = 0.316, d.f. = 2, P = 0.85). Sex did

not affect host-orientation behaviour because there was no

difference in the response of male and female weevils to various

silhouettes for either year (2009: F1,29 = 0.22; P = 0.64; 2010:

χ2 = 4.83, d.f. = 3, P = 0.69).

Experiment 2: does blinding remove host-finding ability

of weevils?

Application of paint to eyes of the insects altered their attraction

to vertical silhouettes. In a preliminary experiment conducted

in 2009, 17 of the 40 weevils originally released into the plots

were recaptured in the pitfall traps. The proportions found in the

(A)

(B)

Figure 2 Total number of weevils of Warren root collar weevils captured

in pitfall traps in each of four silhouette treatments; tree, crown,

trunk and control. Each plot had a treatment trap and an empty

trap, except the control treatment, which had two empty traps.

(A) Experiments were conducted from 25 to 31 August 2009 (n = 40

weevils per treatment) (five insects × four replicates × two trials). An

asterisk indicates a statistically significant difference in proportions

(F1,30 = 5.48; P = 0.0259). (B) Experiments were conducted from 17

to 25 June 2010 (n = 50 weevils per treatment) (five insects × five

replicates × two trials). An asterisk indicates a statistically significant

number of insects trapped in silhouette treatments versus control

(χ2 = 13.7, d.f. = 3, P = 0.004).

treatment versus empty controls differed significantly among

blinded and nonblinded insects (Gadjusted = 5.35, d.f. = 1, P =

0.0207; Fig. 3). Of the 17 weevils recaptured, seven were

blinded and 10 were not. The seven blinded weevils were found

in both empty pitfall traps (n = 3) and those baited with a tree

silhouette (n = 4). By contrast, all 10 of the nonblinded weevils

were found in pitfall traps with silhouettes. This pattern of host- location persisted in the more robust experiment conducted in

2010.

In 2010, only three weevils of the 100 released over the

course of the assays were found in the empty control. Fifty- three of them were found in the traps beneath tree silhouettes.

Insects found in these traps were more frequently nonblinded

(84.9%) versus blinded (15.1%) (Z = −4.52, P < 0.0001;

Fig. 4).

Page 6 of 9

Host-finding behaviour of H. warreni 291

Figure 3 Total number of blinded and nonblinded Warren root collar

weevils captured in empty pitfall traps and pitfall traps with tree

silhouettes. Each plot had a treatment trap and an empty trap, except

the control treatment, which had two empty traps. Experiments were

conducted from 12 to 14 September 2009 (n = 20 weevils per treatment)

(five insects × four replicates × two trials). An asterisk indicates a

statistically significant difference in proportions (Gadjusted = 5.35, d.f. = 1,

P = 0.0207).

Total Number of Weevils Captured

Vision Impaired by Paint Non-Blinded

Figure 4 Total number of Warren root collar weevils with vision impaired

by paint versus nonblinded controls captured in pitfall traps with tree

silhouettes. Experiments were conducted from 11 to 13 June and

6 to 8 July 2010 (n = 50 weevils per treatment) (five insects × five

replicates × two trials). An asterisk indicates a statistical significance

difference from the other treatment (Z = −4.52, P < 0.0001).

Experiment 3: are weevils attracted to colour?

The attraction of weevils to silhouettes did not change when

colours of the trunk and crown portions were manipulated.

Only two out of 54 captured weevils were found in the empty

control (Fig. 5). The insects demonstrated equal attraction to

silhouettes composed of combinations of white/green crowns

and white/black trunks (χ2 = 0.4615, d.f. = 3, P = 0.93).

Discussion

The finding that vision plays an important role in host finding

behaviour by Warren root collar weevil is consistent with

Figure 5 Total number of Warren root collar weevils captured in arena

choice bioassays with a control pitfall trap and a pitfall trap baited

with one of four silhouette treatments: green crown/black trunk, green

crown/white trunk, white crown/black trunk and white crown/white trunk

no significant difference among silhouette treatments; (χ2 = 0.4615,

d.f. = 3, P = 0.93). Experiments were conducted from 3 to 23 August

2010 (n = 40 weevils per treatment) (five insects × two replicates × four

trials).

behaviour observed in several other wood-boring insects that

integrate visual cues of vertical silhouettes with other cues

(Hunt & Raffa, 1991; Strom et al., 1999; De Groot & Nott,

2001; Strom et al., 2001; Goyer et al., 2004; Campbell &

Borden, 2006a, b; Campbell & Borden, 2009; Reeves, 2011)

The strong role of vision in host orientation by Warren root

collar weevil in the absence of any discernible chemical stimuli

to date, however, is rare among root-boring insects. Among

species of Hylobius, attraction to odourless visual stimuli has

only been previously noted in large pine weevil Hylobius

abietis L. (Bjorklund ̈ et al., 2005). In most insects, responses

to olfactory cues play a dominant role in host selection, and

responses to visual cues only occur when appropriate chemical

cues are present (Bernays & Chapman, 1994; Bernays, 2003).

When searching for nondamaged conifer hosts, H. abietis is

equally attracted to traps baited solely with chemicals or solely

with visual stimuli; however, the strongest attraction is to traps

baited with both chemical and visual stimuli (Bjorklund ̈ et al.,

2005). The distinctive response of Warren root collar weevil

adults to visual stimuli in the absence of chemical stimuli

may be explained, in part, by this insect’s range of host plants

and limited dispersal capabilities associated with its life-history

strategy and mode of movement.

Warren root collar weevils are oligophagous, feeding on a

variety of hosts in the Pinaceae family, including species of

Pinus, Picea, Abies, Larix and Tsuga (Warren, 1956; Wood,

1957; Whitney, 1961; Wood & Van Sickle, 1989; Cerezke,

1994; Hopkins et al., 2009). Oligophagous and monophagous

insects tend to be visual specialists compared with polyphagous

insects because plants within the same family are more likely

to have similar morphologies than plants found in different

Page 7 of 9

292 L. A. Machial et al.

families (Prokopy & Owens, 1978). Plants showing similar

morphologies allow the insects that feed on them to develop

specific search images that aid in host location (Prokopy &

Owens, 1978, 1983; Aluja & Prokopy, 1993; Stenberg &

Ericson, 2007).

Visual cues are often important in short range host selection,

from distances of a few centimetres up to 10 m (VanderSar

& Borden, 1977; Bernays, 2003). Movement by adult Warren

root collar weevils falls within this range because the insects

traverse a mean distance of up to 2 m per night (Cerezke, 1994;

Klingenberg et al., 2010a; Machial et al., 2012). Flightless

insects may not be as dependent upon long-distance (i.e.

chemical) cues if the insects typically exist within host pools not

ephemeral in space and time. For example, bark beetles require

olfactory capabilities to process a cacophony of competing host- and nonhost volatiles in flight as they seek either new live trees

(in ‘aggressive’ tree-killing species) or stressed and weakened

hosts (in ‘secondary’ species) (Huber et al., 2000; Raffa, 2001).

By contrast, Warren root collar weevil can persist for more

than one generation on a mature coniferous tree without killing

the host (Cerezke, 1994), and many suitable hosts of various

species are often found together in a coniferous forest. After

using visual stimuli to locate and arrive at a potential host, other

sensory cues, such as tactile and gustatory stimuli, are likely

to be incorporated into the host selection process (Bernays

& Chapman, 1994). Exploitation of chemical cues for host

location is likely critical for larval life stages, similar to

H. abietis (Nordenhem & Nordlander, 1994; Nordlander et al.,

1997) because female Warren root collar weevil tend to be egg

scatterers that deposit their eggs near potential hosts.

Our experiments were not designed to characterize seasonal

variations in responses of adult Warren root collar weevils

to host cues. Seasonal variations can result from changes

in weevil age and reproductive status, causing shifts in

the insects’ biological requirements (Nordenhem & Eidmann,

1991; Hoffman et al., 1997). For example, early in the season,

the insects may be searching for hosts on which to feed before

mating, whereas, late in the season, females may be looking for

hosts near which to oviposit. Declining discrimination as female

behaviour shifts from feeding to egg scattering (Minkenberg

et al., 1992; Cerezke, 1994; Hopkins et al., 2009) is consistent

with our results from 2009 to 2010 (Fig. 2), in which a higher

level of discrimination among silhouettes in 2010 coincided

with an earlier season assay (June 2010 versus August 2009).

We were cautious when making comparisons between years,

however, because the plots were enlarged in 2010. As such,

the pattern in 2010 could simply have reflected shelter-seeking

rather than a feeding or ovipositional strategy. In preliminary

assays, however, we found that the insects would readily

climb both PVC pipes and artificial Christmas trees, comprising

behaviour more indicative of searching for food than avoiding

predators or inclement weather.

Many insects that feed on conifer trees are more attracted

to black silhouettes than to white ones (Dubbel et al., 1985;

Strom et al., 1999; Campbell & Borden, 2006a, b; Campbell

& Borden, 2009), putatively because black silhouettes more

closely resemble the trunks of host trees, whereas white

silhouettes more closely resemble the trunks of nonhost

angiosperms. Throughout all experiments conducted in the

present study, weevils demonstrated a preference for full

tree silhouettes, regardless of black, white or green colour

combinations. The attraction to silhouettes may have masked

weaker responses to colour because we did not directly

compare colours in a choice assay. Alternatively, we may have

failed to test the most optimal colour to which Warren root

collar weevils are attracted. Because the insects are putatively

nocturnal, future research should focus on colour contrast,

spectral sensitivity and silhouette angle.

Currently, there is no easy and accurate method for esti- mating weevil population sizes in forests or plantations. Tree

mortality may lag the appearance of high numbers of adult

Warren root collar weevils by 2 or 3 years because larvae, the

most damaging life stage to young trees, typically take 2 years

to develop to adults (Cerezke, 1994). Furthermore, infestation

levels are typically higher than what is apparent by mortality

rates (Schroff et al., 2006). As a result, high levels of dam- age may occur before it becomes apparent that populations

of Warren root collar weevil are at critical levels. Our results

suggest that a trap could potentially be developed by placing

nondestructive tree-trunk funnel traps that capture live Warren

root collar weevils known as Bjorklund traps around the base ̈

of plastic vertical silhouettes (Bjorklund, 2009). Using such ̈

apparatuses versus live trees could allow easy placement and

transport within forested areas where they would be most effec- tive, such as within harvested sites to reduce weevil numbers

before replanting, or at the margins of young plantations to

reduce weevil ingress (Klingenberg et al., 2010b).

Acknowledgements

We thank the Prince George Tree Improvement Station for

providing research space. We also extend our gratitude to

Kathryn Berry, Gareth Hopkins, Honey-Marie de la Giroday,

Genny Michiel, Rurik Muenter, Dr Kishan Sambaraju and

Robin Steenweg (University of Northern British Columbia)

for their excellent field and laboratory assistance. Funding for

this project was provided by the Canadian Forest Service, the

University of Minnesota College of Agriculture, Food, and

Natural Resources, an NSERC Canada Graduate Scholarship

and a UNBC Research Seed Grant to L.A.M., and an NSERC

Discovery Grant to B.H.A. E. B. Radcliffe (University of

Minnesota), Lynne Rieske-Kinney (University of Kentucky)

and five anonymous reviewers provided valuable comments on

earlier drafts of this manuscript.

References

Aluja, M. & Prokopy, R.J. (1993) Host odor and visual stimulus

interaction during intratree host finding behaviour of Rhagoletis

pomonella flies. Journal of Chemical Ecology, 19, 2671–2696.

Aukema, B.H., Carroll, A.L., Zhu, J., Raffa, K.F., Sickley, T.A. &

Taylor, S.W. (2006) Landscape level analysis of mountain pine

beetle in British Columbia, Canada: spatiotemporal development

and spatial synchrony within the present outbreak. Ecography, 29,

427–441.

Awan, M.S., Wilson, L.T. & Hoffmann, M.P. (1989) Prey location by

Oechalia schellembergii. Entomologia Experimentalis et Applicata,

51, 225–231.

Page 8 of 9

Host-finding behaviour of H. warreni 293

Bernays, E. (2003) Phytophagous insects. Encyclopedia of Insects (ed.

by V. H. Resh and R. T. Carde), pp. 902–905. Academic Press Inc, ́

San Diego, California.

Bernays, E. & Chapman, R.F. (1994) Host-Plant Selection by Phy- tophagous Insects. Chapman & Hall, Inc., New York, New York.

Bjorklund, N. (2009) Non-destructive tree-trunk funnel trap for ̈

capturing Hylobius warreni (Coleoptera: Curculionidae) ascending

stems of trees. Canadian Entomologist, 141, 422–424.

Bjorklund, N., Nordlander, G. & Bylund, H. (2005) Olfactory and ̈

visual stimuli used in orientation to conifer seedlings by the pine

weevil, Hylobius abietis. Physiological Entomology, 30, 225–231.

Borden, J.H., Hunt, D.W.A., Miller, D.R. & Slessor, K.N. (1985)

Orientation in forest Coleoptera: an uncertain outcome of responses

by individual beetles to variable stimuli. Mechanisms in Insect

Olfaction (ed. by T. L. Payne, M. C. Birch and C. E. J. Kennedy),

pp. 97–109. Clarendon Press, U.K.

Campbell, S.A. & Borden, J.H. (2006a) Close-range, in-flight integra- tion of olfactory and visual information by a host-seeking bark beetle.

Entomologia Experimentalis et Applicata, 120, 91–98.

Campbell, S.A. & Borden, J.H. (2006b) Integration of visual and

olfactory cues of hosts and non-hosts by three bark beetles

(Coleoptera: Scolytidae). Ecological Entomology, 31, 437–449.

Campbell, S.A. & Borden, J.H. (2009) Additive and synergistic

integration of multimodal cues of both hosts and non-hosts during

host selection by woodboring insects. Oikos, 118, 553–563.

Carmer, S.G. & Swanson, M.R. (1973) An evaluation of ten pairwise

multiple comparison procedures by Monte Carlo methods. Journal

of the American Statistical Association, 68, 66.

Cerezke, H. (1994) Warren rootcollar weevil, Hylobius warreni Wood

(Coleoptera: Curculionidae), in Canada: ecology, behavior, dam- age relationships, and management. Canadian Entomologist, 126,

1383–1442.

Claver, M.A. & Ambrose, D.P. (2001) Impact of antennectomy, eye

blinding and tibial comb coating on the predatory behaviour of

Rhynocoris kumarii Ambrose and Livingstone (Het., Reduviidae)

on Spodoptera litura Fabr. (Lep., Noctuidae). Journal of Applied

Entomology, 125, 519–525.

De Groot, P. & Nott, R. (2001) Evaluation of traps of six different

designs to capture pine sawyer beetles (Coleoptera: Cerambycidae).

Agricultural and Forest Entomology, 3, 107–111.

Dethier, V.G. (1983) Introduction. Herbivorous Insects: Host-Seeking

Behavior and Mechanisms (ed. by S. Ahmad), pp. 14–16. Academic

Press Inc., New York, New York.

Dubbel, V., Kerck, K., Sohrt, M. & Mangold, S. (1985) Influence

of trap color on the efficiency of bark beetle pheromone traps.

Zeitschrift fuer Angewandte Entomologie (Germany), 99, 59–64.

Duke, L. & Lindgren, B.S. (2006) Attack by Hylobius warreni on

grafted lodgepole pine and its relationships with monoterpene

composition and scion: rootstock diameter ratio. Agricultural and

Forest Entomology, 8, 305–311.

Gotelli, N.J. & Ellison, A.M. (2004) A Primer of Ecological Statistics.

Sinauer Associates Inc., Sunderland, Massachusetts.

Goyer, R.A., Lenhard, G.J. & Strom, B.L. (2004) The influence of

silhouette color and orientation on arrival and emergence of Ips pini

engravers and their predators in loblolly pine. Forest Ecology and

Management, 191, 147–155.

Grant, J. (1966) The hosts and distribution of the root weevils Hylobius

pinicola (Couper) and H. warreni Wood in British Columbia.

Journal of the Entomological Society of British Columbia, 63, 3–5.

Hoffman, G.D., Hunt, D.W.A., Salom, S.M. & Raffa, K.F. (1997)

Reproductive readiness and niche differences affect responses of

conifer root weevils (Coleoptera: Curculionidae) to simulated host

orders. Physiological and Chemical Ecology, 26, 91–100.

Hoover, S. (2000) What do weevils do all night? The adult feeding

and host selection behaviour and diurnal activity of Hylobius

warreni Wood (Coleoptera: Curculionidae). Undergraduate Thesis,

University of Northern British Columbia.

Hopkins, G., Klingenberg, M.D. & Aukema, B.H. (2009) Temptations

of weevil: feeding and ovipositional behaviour of Hylobius warreni

Wood on host and nonhost bark in laboratory bioassays. Agricultural

and Forest Entomology, 11, 397–403.

Horning, J.A. & Lindgren, B.S. (2002) Stand conditions and tree

root dysfunction affecting susceptibility of lodgepole pine, Pinus

contorta var. latifolia engel. to attack by the Warren root collar

weevil, Hylobius warreni. Poster Presentation, IUFRO Working

Party 7.01.02-Tree Resistance to Insects, 10–14 June 2002. Flagstaff,

Arizona.

Huber, D.P.W., Gries, R., Borden, J.H. & Pierce, H.D., Jr (2000) A

survey of antennal responses by five species of coniferophagous bark

beetles (Coleoptera: Scolytidae) to bark volatiles of six species of

angiosperm trees. Chemoecology, 10, 103–113.

Hunt, D.W.A. & Raffa, K.F. (1989) Attraction of Hylobius radicis and

Pachylobius picivorus (Coleoptera: Curculionidae) to ethanol and

turpentine in pitfall traps. Environmental Entomology, 18, 351–355.

Hunt, D.W.A. & Raffa, K.F. (1991) Orientation of Hylobius pales and

Pachylobius picivorus (Coleoptera: Curculionidae) to visual cues.

Great Lakes Entomologist, 24, 225–229.

Klingenberg, M.D., Bjorklund, N. & Aukema, B.H. (2010a) Seeing the ̈

forest through the trees: differential dispersal of Hylobius warreni

within modified forest habitats. Environmental Entomology, 39,

898–906.

Klingenberg, M.D., Lindgren, B.S., Gillingham, M.P. & Aukema, B.H.

(2010b) Management response to one insect pest may increase vul- nerability to another. Journal of Applied Ecology, 47, 566–574.

Kral, K. (2008) Spatial vision in binocular and monocular common

field grasshoppers (Chorthippus brunneus). Physiological Entomol- ogy, 33, 233–237.

Machial, L.A., Lindgren, B.S., Steenweg, R. & Aukema, B.H. (2012)

Dispersal of Warren root collar weevil in three types of habitat.

Environmental Entomology, in press.

Minkenberg, O.P., Tatar, M. & Rosenheim, J.A. (1992) Egg load as

a major source of variability in insect foraging and oviposition

behavior. Oikos, 65, 134.

National Research Council of Canada (2011) Sunrise/Sunset Cal- culator. British Columbia [WWW document]. URL http://www.

nrc-cnrc.gc.ca/eng/services/hia/sunrise-sunset.html [accessed on 4

November 2011].

Nordenhem, H. & Eidmann, H.H. (1991) Response of the pine weevil

Hylobius abietis L. (Col., Curculionidae) to host volatiles in different

phases of its adult life cycle. Journal of Applied Entomology, 112,

353–358.

Nordenhem, H. & Nordlander, G. (1994) Olfactory oriented migration

through soil by root-living Hylobius abietis (L.) larvae (Col.,

Curculionidae). Journal of Applied Entomology, 117, 457–462.

Nordlander, G. (1987) A method for trapping Hylobius abietis (L.) with

a standardized bait and its potential for forecasting seedling damage.

Scandinavian Journal of Forest Research, 2, 199–213.

Nordlander, G., Nordenhem, H. & Bylund, H. (1997) Oviposition pat- terns of the pine weevil Hylobius abietis. Entomologia Experimen- talis et Applicata, 85, 1–9.

Ohrn, P., Klingenberg, M.D., Hopkins, G. & Bj ̈ orklund, N. (2008) ̈

Two non-destructive techniques for determining the sex of live adult

Hylobius warreni. Canadian Entomologist, 140, 617–620.

Prokopy, R.J. & Owens, E.D. (1978) Visual generalist with visual spe- cialist phytophagous insects: host selection behaviour and applica- tion to management. Entomologia Experimentalis et Applicata, 24,

609–620.

Prokopy, R.J. & Owens, E.D. (1983) Visual detection of plants by

herbivorous insects. Annual Review of Entomology, 28, 337–364.

Page 9 of 9

294 L. A. Machial et al.

R Development Core Team (2011) R: A Language and Environment

for Statistical Computing. R Foundation for Statistical Computing,

Austria [WWW document]. URL http://www.R-project.org [accessed

in September 2011].

Raffa, K.F. (2001) Mixed messages across multiple trophic levels:

the ecology of bark beetle chemical communication systems.

Chemoecology, 11, 49–65.

Raffa, K.F. & Hunt, D.W.A. (1988) Use of baited pitfall traps for mon- itoring Pales weevil, Hylobius pales (Coleoptera: Curculionidae).

Great Lakes Entomologist, 21, 123–125.

Reeves, J.L. (2011) Vision should not be overlooked as an important

sensory modality for finding host plants. Environmental Entomology,

40, 855–863.

Reeves, J.L., Lorch, P.D. & Kershner, M.W. (2009) Vision is important

for plant location by the phytophagous aquatic specialist Euhry- chiopsis lecontei Dietz (Coleoptera: Curculionidae). Journal of Insect

Behavior, 22, 54–64.

Rieske, L. & Raffa, K.F. (1993) Potential use of baited pitfall traps in

monitoring pine root weevil, Hylobius pales, Pachylobius picivorus,

and Hylobius radicis (Coleoptera: Curculionidae) populations and

infestation levels. Journal of Economic Entomology, 86, 475–485.

Robert, J.A. & Lindgren, B.S. (2006) Relationships between root form

and growth, stability, and mortality in planted versus naturally

regenerated lodgepole pine in north-central British Columbia.

Canadian Journal of Forest Research, 36, 2642–2653.

Rossel, S. (1986) Binocular spatial localization in the praying mantis.

Journal of Experimental Biology, 120, 265–281.

Schroff, A.Z., Lindgren, B.S. & Gillingham, M.P. (2006) Random

acts of weevil: a spatial analysis of Hylobius warreni attack on

Pinus contorta var. latifolia in the sub-boreal spruce zone of

northern British Columbia. Forest Ecology and Management, 227,

42–49.

Stenberg, J. & Ericson, L. (2007) Visual cues override olfactory cues

in the host-finding process of the monophagous leaf beetle Altica

engstroemi. Entomologia Experimentalis et Applicata, 125, 81–88.

Strom, B.L., Roton, L.M., Goyer, R.A. & Meeker, J.R. (1999) Visual

and semiochemical disruption of host finding in the southern pine

beetle. Ecological Applications, 9, 1028–1038.

Strom, B.L., Goyer, R.A. & Shea, P.J. (2001) Visual and olfactory

disruption of orientation by the western pine beetle to attractant- baited traps. Entomologia Experimentalis et Applicata, 100, 63–67.

Tilles, D.A., Sjodin, K., Nordlander, G. & Eidmann, H.H. (1986a) Syn- ̈

ergism between ethanol and conifer host volatiles as attractants for

the pine weevil, Hylobius abietis (L.) (Coleoptera: Curculionidae).

Journal of Economic Entomology, 79, 970–973.

Tilles, D.A., Nordlander, G., Nordenhem, H., Eidmann, H.H., Wass- gren, A.B. & Bergstrom, G. (1986b) Increased release of host ̈

volatiles from feeding scars: a major cause of field aggregation in

the pine weevil Hylobius abietis (Coleoptera: Curculionidae). Envi- ronmental Entomology, 15, 1040–1054.

Toivonen, R. & Viiri, H. (2006) Adult large pine weevils Hylobius

abietis feed on silver birch Betula pendula even in the presence of

conifer seedlings. Agricultural and Forest Entomology, 8, 121–128.

VanderSar, T.J.D. & Borden, J.H. (1977) Visual orientation of Pissodes

strobi Peck (Coleoptera: Curculionidae) in relation to host selection

behaviour. Canadian Journal of Zoology, 55, 2042–2049.

Walton, A. (2010) BC Ministry of Forest and Range. Provincial-Level

Projection of the Current Mountain Pine Beetle Outbreak, British

Columbia [WWW document]. URL http://www.for.gov.bc.ca/

hre/bcmpb/BCMPB.v7.BeetleProjection.Update.pdf [accessed on 12

March 2011].

Warner, R.E. (1966) A review of the Hylobius of North America, with

a new species injurious to slash pine (Coleoptera: Curculionidae).

The Coleopterists Bulletin, 20, 65–81.

Warren, G.L. (1956) The effect of some site factors on the abundance

of Hypomolyx piceus (Coleoptera: Curculionidae). Ecology, 37,

132–139.

Whitney, R.D. (1961) Root wounds and associated root rots of white

spruce. Forestry Chronicle, 37, 401–411.

Winder, L. (2004) Marking by abrasion or branding and recapturing

carabid beetles in studies of their movement. International Journal

of Pest Management, 50, 161–164.

Wood, S.L. (1957) The North American allies of Hylobius piceus

(De Geer) (Coleoptera: Curculionidae). Canadian Entomologist, 89,

37–43.

Wood, C.S. & Van Sickle, G.A. (1989) Forest Insect and Disease Con- ditions British Columbia and Yukon 1989. Pacific Forest Research

Center, Canadian Forest Service, Environment Canada, British

Columbia.

Accepted 9 January 2012

First published online 22 March 2012