Comparison of baits containing fipronil and sulfluramid for the control of Vespula wasps

Transcription

1 New Zealand Journal of Zoology ISSN: (Print) (Online) Journal homepage: Comparison of baits containing fipronil and sulfluramid for the control of Vespula wasps R. J. Harris & N. D. Etheridge To cite this article: R. J. Harris & N. D. Etheridge (2001) Comparison of baits containing fipronil and sulfluramid for the control of Vespula wasps, New Zealand Journal of Zoology, 28:1, 39-48, DOI: / To link to this article: Published online: 30 Mar Submit your article to this journal Article views: 550 View related articles Citing articles: 17 View citing articles Full Terms & Conditions of access and use can be found at

2 New Zealand Journal of Zoology, 2001, Vol. 28: /01/ $7.00/0 The Royal Society of New Zealand Comparison of baits containing fipronil and sulfluramid for the control of Vespula wasps R. J. HARRIS Landcare Research Private Bag 6 Nelson, New Zealand harrisr@landcare.cri.nz N. D. ETHERIDGE Department of Conservation Nelson Lakes National Park St Arnaud, New Zealand Abstract Fipronil is a new insecticide being trialed in baits for control of introduced social wasps (Vespula spp.). The efficacy of fipronil was compared with sulfluramid (Finitron wasp bait). In a laboratory bioassay, fipronil was faster acting and equally as toxic at concentrations over 1000 times lower than sulfluramid. Doses of 0.2 mg of fipronil active (0.2 g of bait when used at 0.1%) were sufficient to reduce colony activity by 90%. Trials in a 300-ha site in beech forest used a single poison operation, without prefeeding, and bait stations in either a m or m grid. Use of fipronil reduced colony activity within the treated area by 99.7%, whereas sulfluramid had minimal effect on colony activity. Some foragers reinvaded the treated blocks after treatment with fipronil, but wasp densities remained below estimated ecological damage thresholds for vulnerable invertebrates in this habitat. Baiting in lines resulted in variable control of nearby colonies. Despite the low doses of fipronil needed to kill colonies, low bait attractiveness may limit the success of control operations using fipronil in some situations. Keywords baiting Vespula; wasps; fipronil; sulfluramid; Z00008 Received 9 June 2000; accepted 7 September 2000 INTRODUCTION Introduced wasps (Vespula spp.) reach high population densities at many sites in New Zealand (Thomas et al. 1990; Beggs in press). At the peak of the season these high wasp populations can conflict with commercial and recreational users of a site (e.g., Clapperton et al. 1989), and have detrimental impacts on the native fauna (Barr et al. 1996; Toft & Rees 1998; Beggs & Rees 1999; Beggs in press). Biological control options are being pursued (Donovan etal. 1989; Beggs etal. 1996; Rose etal. 1999), but are not yet showing any measurable reduction in wasp populations. The most optimistic estimates suggest that the parasitoid Sphecophaga vesparum could reduce wasp numbers by up to 25% (Toft et al. 1999), much less than the 80% reductions needed to reduce densities below an estimated ecological damage threshold (Toft & Rees 1998; Beggs & Rees 1999). Toxic baiting is the most effective method currently available for reducing wasp numbers by significant amounts, albeit within limited areas. Finitron, a powdered concentrate containing sulfluramid mixed with a protein bait base, is currently registered in New Zealafid for wasp control. This product has been used successfully in beech forest containing honeydew (Spurr 1993; Spurr et al. 1996; Beggs et al. 1998), but has had mixed success in some other habitats (Spurr & Elliott 1996), and has been unable to reduce wasp densities below estimated ecological damage thresholds of 2.7 wasps/ Malaise trap/day for vulnerable Lepidoptera in beech forest (Beggs & Rees 1999), and a similar threshold for spiders (Toft & Rees 1998). Sulfluramid is a slow-acting insecticide, allowing foragers to collect large amounts of bait before dying, and it can take more than 10 days to get a 90% reduction in the activity of colonies (Spurr 1993). A slow-acting toxin may have advantages over a fastacting one if each forager needs to take back large quantities of bait to kill a colony. However, Beggs et al. (1998) reported similar success using either sulfluramid or the fast-acting toxin Wasp Published online 30 Mar 2010

3 40 New Zealand Journal of Zoology, 2001, Vol. 28 wasps in beech forest with honeydew, and discuss factors that may result in poisoning success or failure. STUDY SITES St Arnaud Range Fig. 1 Details of the "Mainland Island" block used for poisoning trials at Lake Rotoiti and inset of northern South Island showing location of the study sites. I, Mainland Island at Lake Rotoiti; 2, Mt Misery; and 3, Gran vi lie State Forest. colony activity was reduced, and fewer foragers were trapped, within poisoned sites. Although 1080 is effective against wasps (Spurr 1991), it is more toxic to vertebrates than insects, so not appropriate for use in wasp baits accessible to birds, domestic animals, and humans (Akre 1991). A fast-acting toxin has the potential advantage of alleviating problems sooner, and allowing rapid assessment of the success of a control operation. The insecticide fipronil has been proposed for use against wasps. Fipronil is a member of the relatively new phenylpyrazole class of neurotoxic insecticides. It blocks neurological inhibition (the gamma amino butyric acid (GABA)-gated chloride channel/ ionopore complex; Scharf & Siegfried 1999). The insecticide shows much greater effectiveness at blocking this complex in insects (LD50 housefly 0.13 mg/kg) than in mammals (LD50 mouse 41 mg/ kg) (Hainzl & Casida 1996). Strategies for the use of fipronil have been developed for a wide range of invertebrate pests from fleas (Siphonaptera) (Atwell et al. 1997) to fire ants (Hymenoptera, Formicidae) (Collins & Callcot 1998). Here we report on laboratory and field trials using the toxins fipronil and sulfluramid for the control of Field trials were all conducted at sites in northwestern South Island (Fig. 1). Sites used had beech trees (Nothofagus spp.) infested with the honeydewproducing scale Ultracoelostoma spp. (Beggs in press). Sites used in the Granville State Forest (S.F.) consisted of open valley floors of levelled mine tailings, with beech forest on the adjacent ridges. Several sites were used within the forest from two access roads (GR NZMS 260 K & NZMS 260 K ; m ASL). A large scale trial was conducted in beech forest in the Nelson Lakes National Park. The treatment sites were within the 825 ha Rotoiti Nature Recovery Project, or "Mainland Island", at St Arnaud on the edge of Lake Rotoiti (GR NZMS 260 N ; m ASL), the site of an intensive pest control programme (Butler 1998). This programme forms part of a national conservation project aimed at releasing native species from predation by introduced pests at key sites on the New Zealand mainland (B. Mansfield unpubl. paper delivered at IUCN World Conservation Congress, Montreal 1996). The nontreatment site was at Mt Misery, an area of beech forest on the edge of Lake Rotoroa (GR NZMS 260 M ; m ASL). METHODS Toxicity in a laboratory bioassay A 20% fipronil suspension concentrate was added to a 30% sugar-syrup to make concentrations of 0.001%, 0.01%, 0.1%, and 1.0% fipronil (wt/vol). A sugar syrup solution was used in the laboratory rather than sardine bait, as is used in the field trials, so a concentration gradient could be prepared and precise doses fed individually to larvae using a pipette. Sugar-syrup without fipronil added was used as a 0% control. Sulfluramid, which does not disperse in water, was dissolved in acetone then mixed with sunflower oil. The acetone was allowed to evaporate before a concentration gradient of 0.001%, 0.01%, 0.1%, and 1.0% (wt/vol) sulfluramid was prepared in a 30% sugar-syrup solution. Sugar-syrup and acetone without sulfluramid added was used as a 0% control.

4 Harris & Etheridge Toxins for wasp control 41 Several nests of Vespula vulgaris were dug from suburban Christchurch and all adults removed. In the laboratory, the wasp comb was broken into sections containing from 33 to 66 late instar larvae. Each comb section formed a replicate (3 replicates per treatment) and wasp larvae were each fed 5 \i\ of the treatment solutions on 29 January This equated to doses of 0, 0.05, 0.5, 5, and 50 ig of active ingredient per larva. The sections of comb containing larvae were maintained at 30 C from 29 January. The total number of larvae entering pupation and emerging as adults was recorded. ANOVA and Fisher's Least Significant Difference Tests (LSD) were used to determine whether there were any significant treatment effects for each insecticide and to compare between the controls for sulfluramid and fipronil. Dosing individual colonies Sardine baits of 0.1, 1, and 2 g containing either 0.1 % or 0.01 % fipronil and non-toxic sardines were presented on small plastic trays directly at the entrance of individual colonies along the edge of Granville Rd (Granville S. F.) on 13 February 1998 (three replicates of each treatment). This corresponded to 6 treatments of 0, 0.01, 0.1, 0.2, 1, and 2 mg of fipronil active ingredient per colony. Previous tests of this method had confirmed that only foragers of the treated colony rapidly discover and collected the bait. Colonies were revisited later the same day to confirm that workers were collecting the bait. The traffic rate of colonies (the number of wasps entering and leaving the nest in three 1-minute counts) was used as an estimate of colony size (Malham et al. 1991). Traffic rate was recorded before, and 6 days after, feeding. Allocation of treatment to colonies was stratified so that each treatment was presented to colonies with a similar range of initial traffic rates. Bait station density Eight sites in Granville S.F. had either 0, 3, 24, or 72 bait stations placed along a 100 m line on the valley floor (two replicates of each). Sites with differing treatments were at least 500 m from each other. On the morning of 12 February 1998, sardine baits containing 0.1% fipronil were placed at each site (beginning at NZST and finishing at NZST). Much more bait was supplied at each bait station than was removed by foragers. About 1200 g of bait was placed at each site with 72 bait stations, <800 grams at each of the two sites with 24 bait stations, and <400 g at each of the two sites with only three bait stations. Baits were left out for about 36 h. On 4-5 occasions during the afternoon of 12 February the number of wasps foraging on baits was recorded by an observer walking along the bait line and making instantaneous counts of the number of foragers feeding on baits. Colonies were located at each site, and the distance was measured to the nearest bait station. The traffic rates of at least five colonies within 100 m of bait stations were monitored at each site the day before and 6 days after poisoning. Several colonies between 100 and 200 m of bait stations were also monitored. ANOVA was used to determine treatment effects on post-poison traffic counts. Site and colony size before treatment were included in the analysis as covariates. Colony count data were square root transformed before analysis to normalise. Large scale efficacy trial The baiting operation was conducted within the Mainland Island. Wasps were poison-baited over about 300 ha of the lower slopes of the St Arnaud Range (Fig. 1). The treatment site was split into two blocks. Block A, 162 ha, had 359 bait stations (2.2 per ha) with spacings about 100 x 50 m, and block B, 134 ha, had 668 bait stations (4.98 per ha) with spacings about 50 x 50 m. A second site, Mt Misery, on the edge of Lake Rotoroa, about 20 km from the treatment blocks, received no poison bait and was monitored as a non-treatment area. Because of the large scale of the field study, the results are believed to be generally applicable to this habitat, but confirmation is desirable through further replication. The density of wasp foragers was indexed using the weekly catches in Malaise traps (10 traps in each of the two blocks of the treatment site, and 10 at the non-treatment site) and converted to daily catch of wasps per trap. Missing values (loss of a weekly trap catch due to trap damage) were estimated by taking the mean catch of the weeks before and after the missing value. The impact of poisoning on individual colonies within the treatment blocks was determined by measuring changes in the traffic rate of colonies before and after poisoning. In 1998, 1% sulfluramid in sardines was used (twice the rate recommended on the label of Finitron, but as used by Beggs et al. 1998). Block A received 278 g of bait per station (616 g per ha). Block B received 130 g of bait per station (648 g per ha). Poisoning was conducted on 7 February 1998, and bait was left out for 15 days. The high volume of bait per station was chosen so that the volume of

5 42 New Zealand Journal of Zoology, 2001, Vol. 28 bait placed out per ha was similar to that presented by Beggs et al. (1998), who used higher densities of bait stations. In 1999, 0.1% fipronil in sardines was used. Because of the low bait take with sulfluramid in 1998 and the higher toxicity of the fipronil, less bait was placed out than in Block A received 60 g of bait per station (133 g per ha). Block B received 30 g of bait per station (150 g per ha). Poisoning was conducted on 4 February 1999, and bait was left out for 5 days. When the bait was collected, the amount removed from a sub-sample of 25 baits in each block was visually estimated compared with weighed samples containing 0, 10, 20, 30, 40, 50 and 60 g of bait. The distance of each Malaise trap from the edge of the treated block (blocks A and B combined), excluding the lake edge, was estimated from the site map. In 1999, the proportion of the total wasps in each Malaise trap caught after poisoning was calculated and correlated with the distance from the edge of the treated block. ANOVA was used to compare these proportions between traps less than and greater than 700 m from the edge of the block. ia JJ60 20 B A 1, 1 : Dose (Mg active) O pupating A emerging :*:::::::: = = < O pupating A emerging { RESULTS Toxicity in a laboratory bioassay A similar percentage of larvae fed 0 and 0.05 g of fipronil entered pupation (Fig 2A). At all higher concentrations few, if any, larvae pupated. No larvae fed fipronil at rates ranging from 0.05 to 50 g/larvae successfully emerged as adults. In comparison, a similar percentage of larvae entered pupation for all sulfluramid concentrations, and only at the highest concentration did significantly fewer wasps emerge from pupation than among the control group (Fig 2B; F = 4.4, d.f. = 4, P = 0.026; LSD (50 g vrs control) = 0.011). This indicates fipronil is faster acting, killing wasps before they enter pupation, and lethal at concentrations at least 1000 times lower than sulfluramid. A lower percentage of wasps in each replicate successfully emerged from the sulfluramid control compared with the fipronil control (Fig. 2; F = 34.6, d.f. = 1, P < 0.01), possibly because they were affected by the sunflower oil and acetone used to dissolve the sulfluramid powder. Dosing individual colonies Wasp foragers rapidly removed all the bait placed in the nest entrance. As the amount of fipronil presented increased, the traffic rate decreased (Fig Dose ([jg active) Fig. 2 Effect of the concentration of A, fipronil and B, sulfluramid on the pupation and emergence success of late instar wasp larvae. Mean ± SE. Both poisons were fed to larvae in a sugar solution. 3). About 0.2 mg of fipronil active was required for a 90% reduction in colony traffic rate. This corresponds to 0.2 g of bait entering a colony at the concentrations used in field trials (0.1%) to get a 90% reduction in traffic rate after 6 days. Bait-station density Forager numbers recorded during instantaneous counts at bait stations during the first day of poisoning were low (Table 1). There was no obvious reduction in the amount of bait on any of the bait lines. Despite the very low visitation rates to baits, the numbers of bait stations significantly affected traffic rates after poisoning (Fig. 4; F = 12.9, d.f. = 3,P< 0.001). Only colonies within 100 m of lines with 24 and 72 bait stations were significantly reduced (LSD, P < 0.05). The reduction for 24 and 72 bait stations were similar, and data were combined to compare

6 Harris & Etheridge Toxins for wasp control S 100 c o CJ ^ 8 CJ? 60 H i 20 f Fipronil (mg active) Fig. 3 Reduction in the traffic rate of wasp colonies due to increasing amount of fipronil in sardine bait fed to individual colonies. Mean ± SE. T 2.0 control 3-baits 24-baits 72-baits Number of bait stations per line Fig. 4 Changes in the traffic rate of wasp colonies within 100 m of bait lincs containing differing numbers of fipronil baits. Back-transformed mean ± SE. Table 1 Forager activity, during the first 6 h of poisoning, on toxic baits placed in a 100m line at Granville State Forest in Site Half Ounce 1 Sullivans Creek Duffers right Duffers left Half Ounce 2 Ridge line Treatment (# of baits at site) Instantaneous counts (mean ± SE) Total wasps on bait line Wasps/bait ± ± ± ± ± ± ± ± ± ± 0.20 with the traffic rates of colonies within 100 m to those that were m away. Colonies between both between 0 and 100 m, and between 100 and 200 m away, had significantly lower post treatment traffic rates than control colonies (Fig. 5; F = 22.9, d.f. = 2, P< 0.001). Large scale efficacy trial Similarnumbers of wasp foragers were caught in the non-treatment site in both seasons (Fig. 6). Before poisoning, the increase in wasp numbers caught in Malaise traps in the Mainland Island lagged behind Mt Misery. In 1998, sulfluramid wasp bait neither reduced the number of foragers present in the treated area by an ecologically significant amount (Fig. 6A), nor markedly reduced wasp colony activity compared with the control (Fig. 7A), at either of the two bait-station densities. At no bait station was all the bait consumed, but the amount removed was not estimated. The ecological damage threshold was

7 44 New Zealand Journal of Zoology, 2001, Vol. 28 M Pre-treatment Post-treatment poisoning (0.50 ± 0.06 wasps caught after poisoning) compared to the catch at traps within 700 m of the edge of the treatment area (0.64 ± 0.03 of wasps caught after poisoning) (F = 5.3, d.f. = 1, P = 0.03). Of the 25 baits in each block scored for bait removal, only one in each block was empty, all others had more than 5 g remaining, and overall about 40% of the bait was removed. control 0 - < 100m m Distance from bait Fig. 5 Effect of distance from bait lines on the reduction in the traffic rate of wasp colonies. Data is from sites with either 24 or 72 baits in 100 m lines. Back-transformed mean + SE. Bars with the same letter above are not significantly different at the 95% level (Fisher's Least Significant Difference Test). exceeded for nearly 3 months. In contrast, very low numbers of wasps were caught in Malaise traps after fipronil was used in 1999 (Fig. 6B): the ecological damage threshold was exceeded only in the sampling period just before poisoning, and approached it on one other sampling date in early April. The traffic rate of all colonies monitored within the poison area was significantly reduced (Fig. 7B). The activity of 16 colonies in Block A, counted the day after poisoning, had dropped by 99.7 % from a mean (± SE) traffic rate of 31.1 ±4.0 wasps/min to 0.1 ± 0.1 wasps/min. The proportion of total wasps caught in Malaise traps that were caught after poisoning showed no significant relationship with the distance of the trap from the edge of the treatment blocks (Fig. 8). However, there is an indication of a reduced catch in traps more than 700 m into the treated block. For these four traps, a significantly lower proportion of the total wasp catch was from the period after the DISCUSSION Fipronil was highly effective in controlling wasps during the large scale (300 ha) operation in All colonies within the treated site were controlled by a single poisoning (99.7% reduction in colony activity), with 2.2 and 5.0 baits/ha. For the first time, we have a technique which will reduce wasp populations enough to protect vulnerable native invertebrates. Fipronil offers major advances for control of wasps compared with sulfluramid (Finitron wasp bait). Wasps receiving a lethal dose die rapidly, as larvae did in the laboratory assay, so less bait is needed and wasp populations will be reduced more quickly to alleviate public nuisance or ecological damage. It is also effective in lower concentrations (looox less), so consumption of equivalent amounts of bait will produce greater reductions in the wasp population and more cost-effective control. As a result, successful reduction in wasp populations may be achieved at some sites where it has previously been difficult to attract sufficient wasps onto protein baits to gain effective control. Sulfluramid was unable to reduce wasp-colony activity significantly at bait station densities comparable with that used for fipronil (even with greater bait available per station). Insufficient bait was collected by foragers to have a lethal effect on colonies before the bait became unattractive. By contrast, kills of between 80 and 100% of colonies were achieved using either sulfluramid or sodium monofluoroacetate (1080) in 30 ha sites in beech forest using bait station densities of about 8/ha (Beggs et al. 1998). Forager activity in the nonpoisoned control sites used by Beggs et al. (1998) was nearly 4 times higher than at Mt Misery in this trial, reflecting the relatively low wasp populations during these trials in 1998 and The high populations preceding poisoning in the study by Beggs et al. (1998) probably increased both forager activity on baits and the amount of bait consumed, which would have increased the effectiveness of the baiting operation using sulfluramid. When wasps are

8 Harris & Etheridge Toxins for wasp control 45 Fig. 6 Number of wasps caught per day in Malaise traps: A, 1997/ 98 treatment sites were poisoned withsulfluramid;b, 1998/99treatment sites were poisoned with fipronil. Mean + SE. Standard errors were calculated from the variation in Malaise trap catches within each block. Daily averages are graphed on the date the trap was cleared. Malaise trap catches from Block A and B within the treated area were similar, so data were combined. The dotted horizontal line represents the estimated ecological damage threshold for detrimental impacts on vulnerable invertebrates (see text). A 25 a, E z O Rotoiti Mt Misery Rotoiti poisoning Nov Dec Jan Feb Mar Apr May B 25 Rotoiti poisoning Nov Dec Feb Mar Apr May

9 46 New Zealand Journal of Zoology, 2001, Vol. 28 pre-treatment H 9-16 days post days post Control 50x50 m Treatment 50x100 m L - ~ - - " ' Distance into treated block (m) Fig. 8 Effect of distance into the treated blocks in the Mainland Island on wasp activity. Data points represent the proportion of all wasps caught in a trap that were caught posl-poisoning. Control 50x50 m 50x100 m Treatment Fig. 7 Changes in traffic rates for colonies monitored within "Mainland Island" compared with non-treatment sites in A, 1998 (sulfluramid treatment); and B, 1999 (fipronil treatment). Mean ± SE. Standard errors were calculated from the variation in traffic rate between nests within each block. already at low densities, a large increase in the density of bait stations might be necessary to get reductions in wasp activity comparable with that achieved with fipronil. Beggs et al. (1998) reduced the abundance of foragers caught in traps by only 55-70%, and wasp activity remained above the ecological damage threshold of 2.7 wasps/trap/day for months, principally because the treated areas were small (30 ha) and hence there was reinvasion of the sites from further afield. There was also reinvasion of foragers into the Mainland Island but, when fipronil was used, insufficient wasps reinvaded to reach the ecological damage threshold. This was probably due to the combination of the highly successful poisoning of nests within (and probably for some distance surrounding) the blocks, and the relatively low wasp densities at the trial sites during the 1997/98 and 1998/99 summers. In addition, the large size of the treated block may reduce the level of invasion, there was some evidence for a reduced catch in Malaise traps near the centre of the block compared with those on the invasion edges. That the reduction was not evident until over 700 m into the poisoned site indicates that poisoning of small blocks will suffer greater edge effects and problems with reinvasion, as encountered by Beggs et al. (1998). Further increasing the size of the treated area may be necessary to get a greater difference in reinvasion after poisoning between the centre and outer regions of the block. If there is substantial reinvasion by foragers from outside the treated area, an addition baiting operation may extend the area in which colonies are controlled and help maintain low wasp populations for the remainder of the season. However, bait attractiveness may be altered if, after the initial reduction in wasp numbers, alternative food increases in abundance. It may be more

10 Harris & Etheridge Toxins for wasp control 47 effective to treat a larger area initially than to repeat the control operation. The higher the density of bait stations, the greater the likelihood of foragers encountering toxic bait and taking a lethal dose back to their colony. Significant reductions in the activity of wasp colonies were recorded within 100 m on a bait line containing 24 fipronil baits, without any pre-feeding, and with very low bait take (< 1 wasp/ bait station recorded during instantaneous counts) in Using bait stations in a line resulted in a large variation in the effect of a treatment on nests within 100 m. This suggests that, for some colonies, few or no foragers encountered the bait stations, while for other colonies foragers collected lethal doses. Baiting in a systematic grid is likely to increase the chances that foragers from most nests encounter the bait. At the bait station densities used in the Mainland Island trial, the furthest a colony would be from a bait station was about 56 m, and baits would be encountered by foragers heading in four different directions from most colonies. Colonies that were m from bait lines were collecting bait and showed significant reductions in activity, but not as much as colonies m away. There was no evidence that control sites from m from baiting lines were effected. Monitoring of nests at different distances from the Mainland Island is planned to determine how far a poisoning effect extends in a uniform habitat after a successful poisoning operation. Only low volumes of fipronil bait need to be placed out for area-wide control. Assuming nest densities averaging 11.2 nests/ha (range 1 to 33 nests/ha) for northern South Island beech forests (Thomas et al. 1990; Beggs in press), only about 2.24 g (range 1 to 6.6 g) of bait per/ha of 0.1% fipronil needs to be consumed for a 90% reduction in colony activity. In the Mainland Island trial, a minimum of 130 g/ha was placed out, of which an estimated 58.5 g was consumed. The more attractive and/or less perishable the bait, the less is wasted, and the more possible it becomes to place out only the ideal bait loading of <7 g/ha. Bait stations spacings of 50 x 50 m and 50 x 100 m were sufficient when using fipronil to kill all colonies within the treated blocks. Further replication is needed with differing wasp densities and in different habitat types before deciding on the ideal bait station spacings. Operations using bait stations at these densities are labour-intensive. Eight people were involved for about eight hours in poisoning the 300 ha of the Mainland Island, and additional time was spent setting up the grid and bait stations before the first season's poisoning. This labour investment will limit the total area that can be controlled. Further trials are needed to improve the technique. For example, we need to determine how far the spacing can be increased to minimise the baiting effort per unit area while still achieving effective control, to determine the effect of changes in wasp abundance on control success and reinvasion from the surrounds. However, we are confident that suppression of wasp populations to low levels in important recreational or ecological sites is achievable. ACKNOWLEDGMENTS Our thanks to Timberlands West Coast Ltd, particularly B. Hawker, for assistance with trials at Granville Forest. J. Rees, R. Toft, V. Hawkey, K. Barton, A. Rose, M. Thomas, M. Maitland, G. Taylor, J. Thorneycroft, D. Peters, L. Peters, R. Mirza, and J. McConochie all provided assistance with baiting operations. D. Butler, P. McCarthy, and J. Beggs were instrumental in the design of the poison operation within the Mainland Island. The Foundation for Research Science and Technology (through contract C09814), Aventis, and the Department of Conservation funded the research. D. Butler, J. Beggs, E. Spurr, A. Austin, D. Morgan, and an anonymous journal referee provided helpful comments on earlier versions of the manuscript. REFERENCES Akre, R. D. 1991: Wasp research: strengths, weaknesses, and future directions - a review. New Zealand Journal of Zoology 18: Atwell, R.; Fitzgerald, M.; Howlett, B.; Jensen, C.; Johnstone, I.; Page, A.; Robatto, I.; Postal, J. M. 1997: The use of topical fipronil in field studies for flea control in domestic dogs. Australian Veterinary Practitioner 27: 184. Barr, K.; Moller, H.; Christmas, E.; Lyver, P.; Beggs, J. 1996: Impacts of introduced common wasps (Vespula vulgaris) on experimentally placed mealworms in a New Zealand beech forest. Oecologia 105: Beggs, J. R. in press: The ecological consequences of social wasps (Vespula spp.) invading an ecosystem that has an abundant carbohydrate resouce. Biological Conservation. Beggs, J. R.; Harris, R. J.; Read, P. E. C. 1996: Invasion success of the wasp parasitoid Sphecophaga vesparum vesparum (Curtis) in New Zealand. New Zealand Journal of Zoology 23: 1-9.

11 48 New Zealand Journal of Zoology, 2001, Vol. 28 Beggs, J. R.; Rees, J. S. 1999: Restructuring of Lepidoptera communities by introduced Vespula wasps in a New Zealand beech forest. Oecologia 119: Beggs, J. R.; Toft, R. J.; Malham, J. P.; Rees, J. S.; Tilley, J. A. V.; Moller, H.; Alspach, P. 1998: The difficulty of reducing introduced wasp (Vespula vulgaris) populations for conservation gains. New Zealand Journal of Ecology 22: Butler, D. J. 1998: Rotoiti nature recovery project: St Arnaud's honeydew beech mainland island strategic plan. Internal report series No. 29.Nelson/ Marlborough Conservancy, Department of Conservation. Clapperton, B. K..; Alspach, P. A.; Moller, H.; Matheson, A. G. 1989: The impact of common and German wasps (Hymenoptera: Vespidae) on the New Zealand beekeeping industry. New Zealand Journal of Zoology 16: Collins, H.L.; Callcot, A. M.A. 1998: Fipronil-an ultralow-dose bait toxicant for control of red imported fire ants (Hymenoptera, Formicidae). Florida Entomologist 81: Donovan, B. J.; Moller, H.; Plunkett, G. M.; Read, P. C.; Tilley, J. A. V. 1989: Release and recovery of the introduced wasp parasitoid, Sphecophaga vesparum (Curtis) (Hymenoptera: Ichneumonidae) in New Zealand. New Zealand Journal of Zoology 16: Hainzl, D.; Casida, J. E. 1996: Fipronil insecticide - novel photochemical desulfinylation with retention of neurotoxicity. Proceedings of the National Academy of Sciences of the United States of America 93: Malham, J. P.; Rees, J. S.; Alspach, P. A.; Moller, H. 1991: Traffic rate as an index of colony size in Vespula wasps. New Zealand Journal of Zoology 18: Rose, E. A. F.; Harris, R. J.; Glare, T. R. 1999: Possible pathogens of social wasps (Hymenoptera: Vespidae) and their potential as biological control agents. New Zealand Journal of Zoology 26: Scharf, M. E.; Siegfried, B. D. 1999: Toxicity and neurophysiological effects of fipronil and fipronil sulfone on the western corn rootworm (Coleoptera: Chrysomelidae). Archives of Insect Biochemistry & Physiology 40: Spurr, E. B. 1991: Reduction of wasp (Hymenoptera: Vespidae) populations by poison-baiting; experimental use of sodium monofluoroacetate (1080) in canned sardine. New Zealand Journal of Zoology 18: Spurr, E. B. 1993: The effectiveness of sulfluramid in sardine bait for control of wasps (Hymenoptera: Vespidae). Proceedings of the New Zealand Plant Protection Conference 46: Spurr, E. B.; Elliott, G. 1996: User trials with Finitron wasp bait. New Zealand Beekeeper 3: Spurr, E. B.; Dew, K. W.; Read. P. E. C.; Elliott, G. 1996: The effectiveness of a sulfluramid concentrate mixed with canned sardine cat-food for control of wasps. Proceedings of the New Zealand Plant Protection Conference 49: Thomas, C. D.; Moller, H.; Plunkett, G. M.; Harris, R. J. 1990: The prevalence of introduced Vespula vulgaris wasps in a New Zealand beech forest community. New Zealand Journal of Ecology 13: Toft, R. J.; Malham, J. P.; Beggs, J. R. 1999: Mortality and emergence pattern of overwintering cocoons of the wasp parasitoid Sphecophaga vesparum vesparum (Hymenoptera: Ichneumonidae) in New Zealand. Environmental Entomology 28: Toft, R. J.; Rees, J. S. 1998: Reducing predation of orbweb spiders by controlling common wasps (Vespula vulgaris) in a New Zealand beech forest. Ecological Entomology 23: