Establishment and Early Effects ofOmphalapion hookeri(Kirby) (Coleoptera: Apionidae) as a Biological Control Agent for Scentless Chamomile,Matricaria perforataMérat (Asteraceae)

Biological Control 14, 85–95 (1999) Article ID bcon.1998.0679, available online at http://www.idealibrary.com on

Establishment and Early Effects of Omphalapion hookeri (Kirby) (Coleoptera: Apionidae) as a Biological Control Agent for Scentless Chamomile, Matricaria perforata Me´rat (Asteraceae) Alec McClay* and Rosemarie De Clerck-Floate† *Alberta Research Council, PO Bag 4000, Vegreville, Alberta T9C 1T4, Canada; and †Agriculture and Agri-Food Canada, Lethbridge Research Centre, PO Box 3000, Lethbridge, Alberta T1J 4B1, Canada E-mail: [email protected] Received March 27, 1998; accepted October 11, 1998

INTRODUCTION Omphalapion hookeri (Kirby) (Coleoptera: Apionidae) (ⴝ Apion hookeri Kirby) was released and established in four provinces in western Canada (British Columbia, Alberta, Saskatchewan, and Manitoba) as a biological control agent for scentless chamomile, Matricaria perforata Me´rat (Asteraceae), a rapidly spreading, introduced, annual weed of disturbed ground. Larvae of this univoltine European weevil feed on developing achenes in the capitula of scentless chamomile. Overall, 74% of releases resulted in establishment; these included both releases of prediapause adults in late summer or fall and postdiapause adults in early summer. In experimental plots in Alberta, O. hookeri persisted for five consecutive growing seasons from an initial release of 38 adults. Populations in these plots showed year-to-year increases of up to 17-fold, and in mass rearing cages net reproductive rates of up to 64-fold were obtained. Up to 17 O. hookeri developed in a single scentless chamomile seed head in the field. In the field plots each individual of O. hookeri completing development reduced seed production in a head by 11.2 seeds; unattacked heads produced an average of 171 seeds. A host choice test confirmed that O. hookeri would not develop in the herbal chamomile, Chamomilla recutita. Adult O. hookeri dispersed up to 350 m between emergence and oviposition. Our results suggest that O. hookeri can be established across the scentless chamomile-infested areas of the Canadian prairies and that it has potential for reducing the weed’s reproductive output. r 1999 Academic Press Key Words: scentless chamomile; Matricaria perforata; Tripleurospermum perforatum; Omphalapion hookeri; Apion hookeri; Chamomilla recutita; biological control; establishment; weeds; dispersal; seed production.

Scentless chamomile, Matricaria perforata Me´rat [⫽ Tripleurospermum perforatum (Me´rat) Wagenitz], Asteraceae, is an annual or short-lived perennial plant of European origin which is becoming an increasingly severe weed problem in cropland and waste areas in western Canada (Douglas et al., 1991, 1992; Woo et al., 1991; Bowes et al., 1994). It occurs widely in the black, grey, and dark brown soil zones of Alberta and Saskatchewan and can spread rapidly because of its profuse seed production, up to 1.8 million seeds m⫺2 (Woo et al., 1991). It occurs in a wide variety of habitats, including annual and perennial crops, pastures, wasteland, roadsides and ditches, urban, and industrial areas (Woo et al., 1991); in agricultural land it is particularly common along slough margins, in field depressions, and in transition areas such as fence lines and right-of-ways (Bowes et al., 1994). It can cause significant problems by competing with crops. In farmers’ fields in Saskatchewan, scentless chamomile at a density of 20 plants m⫺2 in spring wheat caused yield losses ranging from 30 to 80%. Actual densities of scentless chamomile in these fields ranged up to 70 plants m⫺2. Plot experiments indicated that the winter annual form is particularly competitive and that yield losses due to scentless chamomile were greater in moist years (Douglas et al., 1991). Yield of winter wheat was reduced by summer annual scentless chamomile plants only in a moist year, whereas yields were reduced by winter annual plants in a year of normal precipitation (Douglas et al., 1992). In Alberta, scentless chamomile occurs mainly in central and northern areas and causes problems particularly on solonetzic soils and in weak or young stands of forage crops (Cole, 1994). Scentless chamomile is legislated as a noxious weed in Alberta, Saskatchewan, Manitoba, and part of British Colum85

1049-9644/99 $30.00 Copyright r 1999 by Academic Press All rights of reproduction in any form reserved.

86

MCCLAY AND DE CLERCK-FLOATE

TABLE 1 Summary of Mass Rearing of Omphalapion hookeri in Outdoor Field Cages at Vegreville Year

Source of parents

Date placed in cage

1994

Sherwood Park, AB Bayreuth, Germany Progeny from 1994 Progeny from 1995 Field sites (Vegreville, Edmonton)

June 24 July 5 July 6 June 12 June 3–July 8

1995 1996 1997

bia, and as a class 3 secondary noxious weed under the Canada Seeds Act (Woo et al., 1991). Several products are available for chemical control of scentless chamomile in cereals; however, in other crops and in uncultivated land, chemical control options for scentless chamomile are limited or absent. In canola, only one product, clopyralid, is registered for scentless chamomile control. In seedling forage legumes, in hay and grazing land with legumes, and in other crops such as beans, corn, lentils, peas, and potatoes, no chemical control is available (Ali, 1995). Scentless chamomile was proposed as a target for biological control by Peschken et al. (1989), and the first biological control agent screened against it was an achene-feeding weevil, Omphalapion hookeri (Kirby) (Apionidae). In previous literature this was referred to as Apion hookeri Kirby; the subgenus Omphalapion in which it was placed is now treated as a genus (AlonsoZarazaga, 1990; Ehret, 1990). The biology of O. hookeri was described by Freese (1991) and by Peschken and Sawchyn (1993). It is univoltine. Adult females lay eggs into the flower buds of M. perforata when they begin to open in early summer. The larvae feed on the tubular florets and developing achenes, and pupate among the achenes in a cell formed of chewed plant tissue. Adults emerge from the heads in late summer and overwinter, presumably in litter or soil. O. hookeri is widely distributed in Europe and western Asia (Peschken and Sawchyn, 1993). Recently an adventive population was found in Nova Scotia, Canada (Peschken et al., 1993), but it is not known how or when this population was introduced. O. hookeri was approved for release in Canada in 1992, and this paper reports on the releases, establishment, and dispersal of this insect at field sites in western Canada, its population development on experimental plots, and its effects on seed production. MATERIALS AND METHODS

Source of Materials O. hookeri was obtained from the breeding colony studied by D. P. Peschken at the Agriculture Canada Regina Research Station (originally derived from insects collected at Bayreuth, Germany), directly from a

Date first harvested July 27 August 23 July 17 July 22

Number of parents

Number of progeny

Reproductive rate

6 177 111 92 315

597

3.26

2,558 1,316 20,248

23.05 14.30 64.28

field population at Bayreuth, Germany, and from the adventive Nova Scotia population. Mass Rearing O. hookeri was reared every year from 1994 to 1997 at the Alberta Research Council, Vegreville, in a field cage 1.8 ⫻ 3.6 ⫻ 1.8 m containing approximately 25–30 plants of M. perforata in 30-cm-diameter pots. Adults were placed in the cage in early summer when the scentless chamomile plants were in bud stage; numbers, sources, and dates of introduction of adults are shown in Table 1. Heads were clipped as they ripened and before adult O. hookeri emerged (Table 1). Seed heads were allowed to dry for a few days in the laboratory. To speed the collection of adults, the ripe seed heads were then tumbled for a few seconds in a seed scarifier (Forsbergs, Inc., Thief River Falls, MN) to break the achenes loose from the receptacles; the crumbled heads were placed in a screened cage and adults were collected as they emerged. Emerged adults were then placed in 1.7-liter plastic containers with screened ventilation holes, containing bouquets of scentless chamomile foliage, in incubators at 4°C for overwinter storage from mid October to mid May. Host Choice The herbal chamomile, Chamomilla recutita (L.) Rauschert, used in the preparation of teas and essential oils, is the most economically important species closely related to scentless chamomile. Peschken and Sawchyn (1993) found that in no-choice tests O. hookeri would oviposit in buds of C. recutita at a similar rate to that shown on scentless chamomile. However, no adults emerged from C. recutita plants exposed to oviposition by O. hookeri, indicating that this plant will not support complete development of O. hookeri. In 1994, as a further test of this conclusion, 10 plants of C. recutita, in a flowering stage similar to that of the M. perforata plants, were placed in the mass rearing field cage interspersed among the 33 plants of M. perforata. Seed heads of C. recutita were collected separately from those of M. perforata and processed in the same way to detect any emergence of adults.

ESTABLISHMENT OF Omphalapion hookeri ON SCENTLESS CHAMOMILE

Field Plots Six field plots, each 1.8 ⫻ 3.6 m, were set up at the Alberta Research Council, Vegreville, Alberta. These plots were primarily intended as sites for the release and establishment of O. hookeri and of other biological control agents for scentless chamomile as these became available, rather than as a strictly replicated study. It was initially intended that one caged plot would serve as a release site for O. hookeri, and a second plot was also caged to observe the effects of shading on scentless chamomile in the absence of the insect. By 1994 it was apparent that O. hookeri could not be confined in the cages, so sampling was extended to all six plots until 1996 to observe the development of scentless chamomile and O. hookeri populations. The plots were seeded with M. perforata on May 6, 1993. Because of dry conditions and slow germination, 20 additional greenhouse-grown rosettes of M. perforata were transplanted into each plot on June 14, 1993, to bring the plant population close to 100% scentless chamomile cover. On June 16, 1993, two of the plots were covered by walk-in cages 1.8 ⫻ 3.6 ⫻ 1.8 m high screened with 32 mesh-per-inch Saran fabric (Synthetic Industries, Gainesville, GA). Thirty-eight adult O. hookeri collected from the Sherwood Park, Alberta, release site (see Table 2) were released into one of the caged plots (Plot 3) on July 7, 1993. The cage screens were removed from these plots on August 19, 1993 because a heavy infestation of aphids had developed in the cages. Subsequently the cage covers on these two plots were replaced each spring (April 26, 1994 and April 18, 1995) and removed in the fall (September 29, 1994 and September 21, 1995). Plots were not caged in 1996 or 1997. At 3-week intervals during the growing season, seedlings (ⱕ6 true leaves), rosettes (⬎6 true leaves), bolted plants, stems and percentage cover of scentless chamomile were counted on 20 0.1-m2 quadrats within each plot, the predominant flowering stage of plants in the quadrat was determined according to a scale modified from Freese (1991) (Table 3), and visual observations were made to determine the presence of O. hookeri adults on the plants. The total percentage cover of other plant species was also recorded from 1994 onwards. In 1994, 1995, and 1996, any surviving overwintered rosettes were counted separately and tagged to distinguish them from the current season’s rosettes. Each year in late summer, when most seed heads were ripe, 50 stems were collected at random from each plot. These collections were made on September 13, 1993; August 8, 1994; August 15, 1995; and August 21, 1996. The later sampling date in 1993 reflects the later emergence and flowering of scentless chamomile in that year due to the dry conditions in spring. The numbers of flower heads in each of the 4 flowering stages on each stem were recorded. Separately, a

87

sample of 100 heads in each of stages 3 and 4 was collected from each plot. The number of living and dead O. hookeri larvae, pupae, and adults, empty pupal cells, and number and weight of achenes were recorded from each head. In 1995 and 1996, achenes from each head were tested for germination on moist filter paper in petri dishes at a photoperiod of 16 h and day/night temperatures of 25°C/20°C, respectively. In 1995, the diameter of each head (excluding the ray florets) was measured using a vernier caliper, and in 1996 the diameter and height (from the base of the involucre) of each head were measured. Effects on Seed Production The effect of O. hookeri on seed production in the field plots in 1996 was evaluated by analysis of covariance, with number and weight of seeds per head as dependent variables, flowering stage and plot as independent variables, and seed head volume (height ⫻ diameter2 ) and number of O. hookeri per head as covariates. Heads which produced no seed were omitted from the analysis; in many of these seed development had failed due to attack by a phytoplasma pathogen (H. Khadhair, pers. comm.). The effects on seed production in the mass rearing cage were also assessed by multiple regression of number and seed weight on seed head volume and O. hookeri number for 200 stage 4 heads. Field Releases and Monitoring Documented field releases (Fig. 1) were made at 2 sites in British Columbia, 10 sites in Alberta, 53 sites in Saskatchewan, and 4 sites in Manitoba between 1992 and 1996. Dates, locations, and numbers of weevils released are provided for those sites that were monitored for weevil establishment (Table 2). Until 1995, most releases in Alberta were made in early summer using overwintered (postdiapause) adult weevils which were ready to begin ovipositing; in the other provinces, and in Alberta after 1995, most releases were made in late summer using prediapause adults. In Saskatchewan, 19 recorded releases were made between 1992 and 1994, and 55 were made in 1995. Most releases made in Saskatchewan in 1994, using weevils from Nova Scotia, were not documented. However, it is known that about 10,000 weevils were shipped to Saskatchewan (Glen Sampson, Nova Scotia Agricultural College, pers. comm.) and were divided into lots of 500 for release. All O. hookeri released in 1995–1996 in Saskatchewan were from Nova Scotia. In 1997, fourteen additional releases of 287–1000 prediapause adults were made in Alberta and two releases of 450 prediapause adults were made in Manitoba; these have not yet been monitored for establishment. In Alberta, sites were monitored in early summer

88

MCCLAY AND DE CLERCK-FLOATE

FIG. 1. Release sites of Omphalapion hookeri in western Canada, 1992–1995. Filled circles: successful establishment (overwinter survival and breeding observed at least in the year after release). Open circles: no establishment. Circles with enclosed square: successful and unsuccessful releases in same area.

(June–mid July) by searching for adult weevils on the plants, and in late summer (August–September) by collecting and dissecting samples of 100 or 200 ripe heads from each location and dissecting to determine numbers of each life stage of O. hookeri present. Heads from which adult O. hookeri had already emerged could be easily identified by the presence of an empty pupation cell or cells surrounded by chewed achene tissue. Sites in British Columbia, Saskatchewan, and Manitoba were monitored by collecting 300 ripe seed heads on average (40–1200 depending on availability of plants) from each location in August or September. Some Saskatchewan sites were also visited in spring 1996 and checked for adults. At most sites, the seed heads were collected within 10 m of the original release point, but at Fort St. John (site 2), seed heads were collected near the release point (⬍10 m and 10–20 m), and at a control site 100 m away. Dispersal Between July 25 and 29, 1996, patches of scentless chamomile in the vicinity of the release plots at Vegreville were searched for adult O. hookeri. At this time of year adults are easily visible on the flower buds. Locations of the plants and O. hookeri were plotted using a portable differential GPS unit (CMT PC-5, Corvallis Microtechnology). The same method was used on July 26, 1996, and on July 21, 1997, to plot dispersal from the Terwillegar Park release site in Edmonton.

RESULTS AND DISCUSSION

Mass Rearing Results of the mass rearing from 1994 to 1997 are provided (Table 1). The lower reproductive rate in 1994 was probably because most of the adults used for rearing in this year were collected from the field in Germany relatively late in the season, and females may already have laid many of their eggs before they were collected. The results from 1995 to 1997 indicate that rearing of O. hookeri on potted scentless chamomile plants in outdoor field cages is a simple and effective way to produce large numbers of the insect for field releases. Up to 30 adult O. hookeri were found to have completed development in a single scentless chamomile flower head in the mass rearing cage. Survival of O. hookeri in overwinter storage was very variable. From 1994 to 1995, 37% of adults placed in storage were recovered alive, while from 1995 to 1996 survival was only 6%. Further work is currently under way to determine optimum overwinter storage conditions for O. hookeri. Host Choice Only one adult O. hookeri emerged from flower heads of C. recutita in 1994. Assuming that this individual had actually completed development on C. recutita and was not a stray, the reproductive success on this plant was only 0.17% of that on scentless chamomile, confirm

TABLE 2 Releases and Recoveries of Omphalapion hookeri against Matricaria perforata in Western Canada

Locality

Date of release

Years recovered b Pre- or Number released Source a post-diapause 93 94 95 96 97

Site description and remarks

British Columbia c Fort St. John (56°148N, 120°558W) Site 1 Site 2 Site 2

August 28, 1992 August 28, 1992

53 116

G NS

Pre Pre

N Y

— Y

N Y

— —

Y

Y

—

— Edge of wheat field Y Poplar bluff in center of field Y Same site as previous

August 25, 1993

200

NS

Pre

September 26, 1996

200

G

Pre

N

September 12, 1996

200

G

Pre

Y

Pasture

NS

Post

Y

Park Industrial area (clay storage mound)

Alberta Bezanson (55°138N, 118°288W) Drayton Valley (53°058N, 114°478W) Edmonton (53°338N, 113°288W) Edmonton (53°358N, 113°378W) Grande Prairie (55°128N, 118°288W) Hillcrest (49°358N, 114°238W) Nisku (53°208N, 113°328W) Sherwood Park (53°318N, 113°198W) Vegreville (53°308N, 112°038W) Viking (53°068N, 111°468W)

June 29, 1995

47

Y

October 3, 1996

200

G

Pre

Y

September 26, 1996

200

G

Pre

Y

June 3, 1993

18

R

Post

June 16, 1995

44

NS

Post

June 30, 1992

105

R

Post

38

R NS

July 7, 1993 August 25, 1993

200

N

Y

—

—

N

Y

Roadside along river, near town — Roadside

Y

N

—

— Park, caged release

Post

Y

Y

Y

Y

Pre

N

N

—

Experimental plots, caged release — Waste ground, heavily grazed and trampled by cattle in spring 1994

Saskatchewan d Aylsham (53°128N, 103°498W) Site 1

August 24, 1995

c. 500

NS

Pre

Site 2 Balgonie (50°298N, 104°168W) Bethune (50°438N, 105°138W) Site 1 Site 2

September 6, 1996 August 25, 1993

c. 500 200

NS NS

Pre Pre

July 30, 1995 September, 1996

c. 500 c. 500

NS NS

Pre Pre

Y

Y Y

Broadview (50°228N, 102°358W) Canwood (53°228N, 106°368W) Site 1 Site 2

September 1, 1995

c. 500

NS

Pre

—

N

August 31, 1995 August 31, 1995

c. 500 c. 500

NS NS

Pre Pre

Y Y

Site 3

August 31, 1995

c. 500

NS

Pre

N

Site 4

August 31, 1995

c. 500

NS

Pre

Y

Fall, 1995

c. 500

NS

Pre

Y

Y Gravel pit — Roadside, edge of field, site plowed in 1997 Y Roadside, field edge, 11% heads attacked 1997 — Roadside, edge of field, site mowed in 1997 Y Edge of field/slough

Fall, 1995 Fall, 1995

c. 1000 c. 500

NS NS

Pre Pre

— —

N N

Dubuc (50°418N, 102°288W) Edgeley (50°388N, 103°598W) Site 1 Site 2

89

Y

Y

Y

Y

— Edge of slough/field, plowed in 1997 Y Farm yard, along road — Edge of hay field, site plowed in 1997

Near slough, along road Alfalfa field, none in seed heads, adults seen in 1997 Near farm yard, along pasture

Edge of slough/field Edge of slough/field

TABLE 2—Continued

Locality Esterhazy (50°398N, 102°058W) Site 1 Site 2 Hafford (52°438N, 107°218W) Site 1

Date of release

Years recovered b Pre- or Number released Source a post-diapause 93 94 95 96 97

Fall, 1995 September 1, 1995

c. 500 500

NS NS

Pre Pre

August 1996

c. 500

NS

Pre

August 1996

c. 500

NS

Pre

August 29, 1995

c. 500

NS

Pre

Y

In grass field, 5% heads attacked 1997 Y In grass field, 6% heads attacked 1997 — Site flooded in 1997

August 1, 1995

c. 500

NS

Pre

N

N

c. 1000

NS

Pre

Y

September 15, 1995

500

NS

Pre

N

August 27, 1994

400

NS

Pre

—

Y

August 27, 1994 Fall, 1996

400 c. 500

NS NS

Pre Pre

—

—

September 1, 1995 September 1, 1995 September 1, 1995

c. 500 c. 500 c. 500

NS NS NS

Pre Pre Pre

August, 1994

c. 500

NS

Pre

August, 1994

c. 500

NS

Pre

September 2, 1996

c. 500

NS

Pre

September 1, 1995

c. 500

NS

Pre

Y

— Abandoned farm yard, none in seed heads but adults seen July 1996

September 1, 1995

c. 500

NS

Pre

Y

Y

Site 2 St. Walburg (53°398N, 109°128W)

September 2, 1996 June 29, 1992

c. 500 c. 700

NS R

Pre N/A

—

Y Y

Stockholm (50°398N, 102°188W)

Fall, 1995

c. 400

NS

Pre

Y

Y

Site 2 Hendon (52°058N, 103°508W) Holdfast (50°588N, 105°258W) Indian Head (50°328N, 103°408W)

Kelvington (52°108N, 103°328W) Lloydminster (53°178N, 110°008W) Site 1

Site 2 MacNutt (51°058N, 101°368W) McLean (50°318N, 104°048W) Site 1 Site 2 Site 3 Melville (50°558N, 102°488W) Site 1

Site 2

Moosomin (50°088N, 101°408W) Qu’Appelle (50°338N, 103°538W) Rocanville (50°238N, 101°428W) Site 1

September 1, 1995

90

Y Y

Y N

Site description and remarks

17% heads attacked 1997 Roadside

Y

N

—

Low area in middle of field N Waste site near grain elevator, mowed in 1996 adults seen July 1996 — Part of site bulldozed 1996, site flooded in 1997

— Slough, cultivated field edge, 5% heads attacked in 1996, no plants in 1997 N In ditch along road Y Edge of barley field

N Y N

N N N

Y

Y

N

N

— Edge of cultivated field, sprayed with herbicide in 1997, adults seen July 1996 — Edge of cultivated field, sprayed with herbicide in 1997 Y Along railway

—

Roadside Edge of pasture Waste area along fence

Roadside, 15% heads attacked 1996 Roadside In ditch on alfalfa field edge, potted chamomile with O. hookeri eggs in flower heads transplanted into stand, 10% heads attacked 1997 Edge of slough

TABLE 2—Continued

Locality Tisdale (52°518N, 104°038W) Site 1

Date of release

Years recovered b Pre- or Number released Source a post-diapause 93 94 95 96 97

Site description and remarks

August 24, 1995

c. 500

NS

Pre

Y

Y

August 24, 1995 September 8, 1996 September 9, 1996

c. 500 c. 500 c. 500

NS NS NS

Pre Pre Pre

N

Y N Y

September 1, 1995

c. 500

NS

Pre

Y

September 2, 1996

c. 500

NS

Pre

— Edge of oat field, site flooded 1997, 6% heads attacked 1996 Y Edge of slough

September 14, 1995

c. 100

NS

Pre

N

Y

September 14, 1995

c. 100

NS

Pre

Y

September 14, 1995 September 14, 1995

c. 100 c. 100

NS NS

Pre Pre

N Y

In windbreak hedge, field edge — In windbreak hedge, field edge N Field edge N Field edge

September 1, 1995

c. 500

NS

Pre

Y

Y

Site 2

September 1, 1995

c. 500

NS

Pre

Y

Site 3

September 2, 1996

c. 500

NS

Pre

Site 4

September 2, 1996

c. 500

NS

Pre

September 1, 1995

c. 500

NS

Pre

Y

August 24, 1995

c. 500

NS

Pre

N

— Edge of wheat field, low area

August, 1994

c. 500

NS

Pre

Y

—

Fall, 1996

c. 500

NS

Pre

— Edge of cultivated field, no scentless chamomile at site in 1996 Y Edge of wheat field

August, 1994

2,100

NS

Pre

Y

Y

Y

August, 1992

300

NS

Pre

N

—

—

—

August, 1994 July 2, 1992

500 57

NS G

Pre Pre

Y

—

— —

Y —

— Edge of hay field, accidentally mowed in 1993 N Field edge — Native pasture, gravelly soil, most plants dead in 1993 due to drought

Site 2 Site 3 Site 4

Wapella (50°168N, 101°588W) Site 1

Site 2 Watrous (51°408N, 105°288W) Site 1 Site 2 Site 3 Site 4 Whitewood (50°208N, 102°168W) Site 1

Wishart (51°338N, 103°598W) Wynyard (51°468N, 104°118W) Yorkton (51°138N, 102°288W) Site 1

Site 2

Edge of slough/field, 13% heads attacked 1997 Along ditch of farm road Edge of field Edge of field/farm yard, 6% heads attacked 1997

Roadside, slough edge, 15% heads attacked 1997 Y Edge of field, 16% heads attacked 1997 Y In ditch beside crop, 7.5% heads attacked 1997 N Edge of slough, next to caraway crop, few plants in 1997 — Waste area

Manitoba Dunnotar (50°278N, 96°578W) Erickson (50°308N, 99°558W) Site 1 Site 2 Winnipeg Beach (50°308N, 96°588W)

Low site on edge of lagoon

Note. All were open releases unless shown as caged under ‘‘Site description and remarks.’’ G: Germany; R: Laboratory colony at Regina, derived from Germany; NS: Nova Scotia. b Y, recovery of O. hookeri; N, site monitored but no recovery; —, site not monitored. c Sites in British Columbia were not checked in 1996 due to inclement weather. d At sites in Saskatchewan where a recovery year is shown but the percentage attack is not indicated, attack was less than 5% of seed heads dissected. a

91

92

MCCLAY AND DE CLERCK-FLOATE

TABLE 3 Flowering Stages of Scentless Chamomile; Modified from Freese (1991) Stage

Description

1

Closed or open buds; ray flowers folded in or erect Ray flowers spread out flat; disk flowers less than 50% open; disk flat Ray flowers spread or withering; disk flowers mostly or all open but corollas still attached, disk swelling, no ripe seed Ray flowers withered or fallen, seed beginning to ripen or ripe, corollas of disk flowers falling or fallen

2 3

4

Equivalent stages (Freese, 1991) I–III IV–V VI–VII

VIII–X

ing the conclusions of Peschken and Sawchyn (1993) that O. hookeri is virtually monophagous on scentless chamomile. Experimental Plots 1993–1996 The interpretation of results in the experimental plots was complicated by large between-plot variation and year-to-year changes in scentless chamomile density and flower production (Fig. 2). Scentless chamomile densities declined greatly over the 4 years of the study; this was not an effect of O. hookeri but was primarily due to competition from other plant species, mainly smooth brome, Bromus inermis Leyss. (Poaceae), which progressively invaded the plots. The shading effect of the screen cages on plots 3 and 4 further reduced the vigor of scentless chamomile on these plots, leading to lower flower production and more rapid displacement by B. inermis. In 1996, to ensure that scentless chamomile populations would persist on the

FIG. 2. Flower head production of scentless chamomile (above) and percentage cover of scentless chamomile and other plant species (below) in experimental plots at Vegreville, 1993–1996.

FIG. 3. Omphalapion hookeri population (individuals per m2 ) and percentage of scentless chamomile heads attacked by O. hookeri in experimental plots at Vegreville, 1993–1996.

plots, dead B. inermis top growth was hand weeded from three plots and new growth was sprayed with a selective grass-killing herbicide, Select (clethodim). O. hookeri was recovered in the experimental plots at Vegreville every year from 1993 to 1997. In 1993 sampling was conducted only in the caged plot in which O. hookeri was released. In 1994 it became apparent that O. hookeri was not confined to the caged plot, so from 1994 to 1996 all six plots were sampled for the weevil. Estimated numbers completing development ranged from 67 to 767 adults m⫺2 (Fig. 3). In 1995, estimated O. hookeri populations were lower than in 1994, but rates of attack were higher (Fig. 3); this was because the density of scentless chamomile plants in the plots and their flower head production had declined due to competition from other plant species. The reduced total population in 1995 may have been partly due to dispersal away from the plots. In 1996, the estimated population increased slightly but the rate of attack (weevils per head) increased sharply; this is related to the reduced flower head production in that year (Fig. 2). Densities of O. hookeri in stage 3 heads were usually lower than in stage 4 heads (Fig. 4). As oviposition only occurs into heads in stages 1 and 2, this probably

FIG. 4. Number of Omphalapion hookeri per head of scentless chamomile in flowering stages 3 and 4 in experimental plots at Vegreville, 1993–1996.

ESTABLISHMENT OF Omphalapion hookeri ON SCENTLESS CHAMOMILE

TABLE 4 Effects of Seed Head Volume and Omphalapion Infestation Level on Achene Production per Head of Scentless Chamomile, Vegreville, 1996 Regression coefficients (estimates ⫾ std. error) for Data source Field plots

Rearing cage

Dependent variable

Seed head volume (mm3 ) a

O. hookeri number a

Achene number 0.32 ⫾ 0.011*** ⫺11.18 ⫾ 1.02*** Achene weight (mg) 0.0119 ⫾ 0.0003*** ⫺0.263 ⫾ 0.027*** Achene number 0.17 ⫾ 0.02*** ⫺2.48 ⫾ 0.79** Achene weight (mg) 0.0076 ⫾ 0.0006*** ⫺0.101 ⫾ 0.029**

a

Superscripts indicate estimates significantly different from zero: ** 0.001 ⬍ P ⬍ 0.01; *** P ⬍ 0.001.

reflects the fact that heads which had reached stage 4 by the sampling date were available for oviposition earlier in the season than heads which had reached stage 3, and were thus exposed to higher rates of oviposition. O. hookeri showed a clumped distribution over available flower heads. The variance/mean ratio in the field plots in 1996 was 3.26, consistent with the values found by Freese (1991). The maximum number of adult O. hookeri found in a single head in the field plots was 12. Effects on Seed Production In the field plots, both seed weight and seed number showed significant differences among plots and flowering stages, a highly significant positive regression on seed head volume, and a highly significant negative regression on numbers of O. hookeri (regression coefficients shown in Table 4). Heads attacked by O. hookeri therefore produced fewer seeds than would have been expected from unattacked heads of the same size and flowering stage in the same plot. The regression coefficients indicate that seed production in the field plots was reduced by about 11.2 seeds or 0.263 mg by each O. hookeri developing in a head. In the rearing cage the effects were less marked, although still significant; each O. hookeri reduced seed production by 2.48 seeds or 0.101 mg. The reasons for the differences between the effects in the field plots and in the rearing cage are unknown. They can be only partially explained by larger achene size in the rearing cage; mean individual achene weight, calculated from a regression of achene weights against numbers, was 0.0386 mg in the rearing cage and 0.0297 mg in the field plots. Possibly the potted plants in the rearing cage, being well watered

93

and fertilized and subject to less competition than plants in the field, were better able to compensate for damage by O. hookeri. Both values are considerably lower than the 21 achenes per weevil reported to be needed for complete development by an O. hookeri larva (Freese 1991). Mean germination of seeds from the field plots was 56.5 and 52.9% in 1995 and 1996, respectively. There was no effect of level of O. hookeri infestation on germination in either year (analysis of covariance: 1995 P ⫽ 0.428, 1996 P ⫽ 0.676), indicating that larval feeding by O. hookeri has no detrimental effect on the remaining seeds in the head. Field Releases and Recoveries O. hookeri established successfully at 74% of sites in western Canada where it was released and subsequently monitored (Table 2). Successful establishment occurred throughout most of the geographic range of scentless chamomile on the Canadian prairies (Fig. 1), from southeastern Manitoba to Fort St. John, northeastern British Columbia, which was the highest latitude at which the weevil was released (56.14°N, 120.55°W). A G test (Sokal and Rohlf, 1981) showed no significant differences in frequency of establishment between adults released pre- and postdiapause (73 versus 80%, respectively) or between weevils originating from Germany and from Nova Scotia (70 versus 75%, respectively). Establishment was achieved at some sites with quite small initial releases (e.g., 38 adults at Vegreville, Alberta; 57 adults at Winnipeg Beach, Manitoba). A number of the failures to establish were associated with environmental changes or human activities at the release site, such as drought, mowing, construction, herbicide spraying, or trampling by cattle (e.g., Viking, Alberta; Kelvington, Saskatchewan). Monitoring at some sites was made more difficult by the tendency of dense scentless chamomile infestations to decline as they are replaced by perennial plant species. This was noted in the field plots at Vegreville and at several field release sites such as Sherwood Park and Nisku, Alberta, and Yorkton, Saskatchewan (Table 2). Percentage of seed heads attacked by O. hookeri was initially less than 5% at most sites where the weevil established in Saskatchewan and Manitoba. However, weevil attack was as high as 15% one year after release in SE Saskatchewan (Rocanville; Table 2) and up to 8 weevils were found per head in this region (Rocanville and Whitewood, site 2). At Canwood site 3 (1995 release) up to 17 adult O. hookeri were found in one seed head in 1997. Recovery in consecutive years at a site (e.g. Balgonie; Table 2) did not always show an increase in the number of weevils recovered. However, at other sites monitoring in consecutive years showed increased rates of attack or confirmed the presence of

94

MCCLAY AND DE CLERCK-FLOATE

O. hookeri at sites at which it had not previously been recovered. O. hookeri successfully established both from spring releases of postdiapause adults which had been kept in cold storage over winter and from late summer or fall releases of newly emerged, prediapause adults. Since optimum storage conditions for reliable survival of O. hookeri over winter have not been determined, it may be more advisable at present to use prediapause adults for routine field releases. Dispersal In early summer of 1996, in the fourth year after release, adult O. hookeri were found in small numbers on almost all patches of M. perforata up to 100 m from the Vegreville release plots. Scattered adults were found up to 240 m away from the plots. In 1997, scattered adults were found up to 800 m from the release point. At Terwillegar Park, Edmonton, 1 year after release, adults were found up to 42 m away from the release point, but not on two isolated patches of M. perforata about 75 m from the release point. By 2 years after release, however, O. hookeri was abundant on these two patches, and scattered individuals were found up to 700 m from the release point. At Fort St. John, BC, seed head dissection data showed that within 1 year of release, the weevil had dispersed more than 100 m, including about 50 m across an area devoid of scentless chamomile. As the insect is univoltine, these observations indicate that some individual adults can disperse at least 350 m before ovipositing, and that open areas up to 50 m wide without stands of M. perforata do not seem to represent a barrier to dispersal. It is not known whether this dispersal occurs primarily before or after overwintering. Phenology O. hookeri is univoltine in Alberta, as was found by Freese (1991) in Germany. At the field sites overwintered adults were seen on scentless chamomile plants from early May onwards. Oviposition began as soon as flower buds in a suitable stage were available. In the mass rearing cages, adults were found ready to emerge from the heads on August 5, 1994, and emerged adults were found August 12, 1994. Assuming that the first adults found in heads were the progeny of the six adults placed in the cage on June 24, development from egg to adult was complete in not more than 42 days. In the field plots, adults which were probably newly emerged were seen from early August onwards. General Discussion Our results indicate that it should be possible to establish O. hookeri throughout the areas infested by scentless chamomile in western Canada. The weevil

does not appear to be limited to any particular habitat type, nor does climate appear to be a limiting factor for its establishment within this area. The ultimate effects of this insect will depend on whether it can reach sufficiently high densities to destroy most of the plant’s seed production. In our field plots unattacked scentless chamomile plants produced an average of 171 seeds per head in 1996. Based on the results above, a population density of about 15 weevils per head would be needed to approach complete seed destruction. The maximum number of O. hookeri found in a single head in the field plots was 17; however, up to 30 were found in a head in the mass rearing cages. No parasitoids were reared from any seed heads in the experimental plots or field release sites. This contrasts with the situation in Europe where five parasitoid species are recorded (Freese and Gu¨nther, 1991) and third-instar mortality, mainly from parasitism, can be around 40% (Freese, 1991). Further population monitoring should reveal whether this freedom from parasitism will allow O. hookeri populations in Canada to increase to a level which will cause significant seed destruction in the field. ACKNOWLEDGMENTS We thank Robert B. Hughes, Irving Switzer, Leslie Dietz, Kim Stromme, Sophie Verzosa, and Rick Weste for technical assistance; Diether Peschken, Glen Sampson, Dieter Schroeder, and Andrea Freese for providing collections of O. hookeri; and Chuck Richardson, Russ Jasman, and Christopher Saunders for arranging access to field sites in Alberta. Thanks also to the following people for collections of seed heads for dissection and for information on releases: Ken Nickel and Jill Copes (BC); Eric Johnson, Barb Gradin, Doug Billett, Len Juras, Brad Thompson, Ed Tanner, James Donovan, Judy McKell, Wendy Schatz, Darlene Fisk, Brian Harris, Lewis Reeve, Ernie Patrick, Donald Dyck, Ivan Allin, Leroy Bader, and Doug Patience (SK); and Carla Pouteau, Risa Kennedy, Rick Hnatiuk, and Neil Galbraith (MB). Financial support was provided by the CanadaAlberta Environmentally Sustainable Agriculture Agreement (project RES 048-93) and by Nova Gas Transmission Limited.

REFERENCES Ali, S., Ed. 1995. ‘‘Crop Protection with Chemicals, 1995.’’ Alberta Agriculture, Food and Rural Development, Edmonton, Alberta. Alonso-Zarazaga, M. A. 1990. Revision of the supraspecific taxa in the Palaearctic Apionidae Schoenherr, 1823 (Coleoptera, Curculionoidea). 2. Subfamily Apioninae Schoenherr, 1823: introduction, keys and descriptions. Graellsia 46, 19–156. Bowes, G. G., Spurr, D. T., Thomas, A. G., Peschken, D. P., and Douglas, D. W. 1994. Habitats occupied by scentless chamomile (Matricaria perforata Me´rat) in Saskatchewan. Can. J. Plant Sci. 74, 383–386. Cole, D. E. 1994. ‘‘Scentless Chamomile: Biology and Control.’’ Alberta Agriculture, Food and Rural Development, Edmonton, Alberta. Douglas, D. W., Thomas, A. G., Peschken, D. P., Bowes, G. G., and Derksen, D. A. 1991. Effects of summer and winter annual scentless chamomile (Matricaria perforata Me´rat) interference on spring wheat yield. Can. J. Plant Sci. 71, 841–850.

ESTABLISHMENT OF Omphalapion hookeri ON SCENTLESS CHAMOMILE Douglas, D. W., Thomas, A. G., Peschken, D. P., Bowes, G. G., and Derksen, D. A. 1992. Scentless chamomile (Matricaria perforata Me´rat) interference in winter wheat. Can. J. Plant Sci. 72, 1383–1387. Ehret, J.-M. 1990. Les Apions de France: cle´s d’identification commente´es (Coleoptera Curculionidae Apioninae). Bull. Mens. Soc. Linn. Lyon 59, 209–292. Freese, A. 1991. Apion hookeri Kirby (Col., Curculionidae), a potential agent for the biological control of Tripleurospermum perforatum (Me´rat) Wagenitz [⫽T. inodorum (L.) C.H. Schultz, Matricaria perforata Me´rat, M. inodora L.] (Asteraceae, Anthemideae) in Canada. J. Appl. Entomol. 112, 76–88. Freese, A., and Gu¨nther, W. 1991. The insect complex associated with Tripleurospermum perforatum (Asteraceae: Anthemideae). Entomol. Gen. 16, 53–68. Peschken, D. P., and Sawchyn, K. C. 1993. Host specificity and suitability of Apion hookeri Kirby (Coleoptera: Curculionidae), a candidate for the biological control of scentless chamomile, Matri-

95

caria perforata Me´rat (Asteraceae) in Canada. Can. Entomol. 125, 619–628. Peschken, D. P., Sawchyn, K. C., and Bright, D. E. 1993. First record of Apion hookeri Kirby (Coleoptera: Curculionidae) in North America. Can. Entomol. 125, 629–631. Peschken, D. P., Thomas, A. G., Bowes, G. G., and Douglas, D. W. 1989. Scentless chamomile (Matricaria perforata)—a new target weed for biological control. In ‘‘Proceedings of the VII International Symposium on Biological Control of Weeds, March 6–11, 1988, Rome, Italy’’ (E. S. DelFosse, Ed.), pp. 411–416. Ist. Sper. Patol. Veg. (MAF). Rome. Sokal, R. R., and Rohlf, F. J. 1981. ‘‘Biometry: The Principles and Practice of Statistics in Biological Research,’’ 2nd ed. Freeman, New York. Woo, S. L., Thomas, A. G., Peschken, D. P., Bowes, G. G., Douglas, D. W., Harms, V. W., and McClay, A. S. 1991. The biology of Canadian weeds. 99. Matricaria perforata Me´rat (Asteraceae). Can. J. Plant Sci. 71, 1101–1119.