Field evaluation of para-aminopropiophenone (PAPP) for controlling stoats (Mustela erminea) in New Zealand

Abstract

Stoats (Mustela erminea) have been implicated in the decline of many native species in New Zealand. At present, kill-trapping is the main technique used for controlling them, but it is labour intensive and expensive. Poisoning offers an alternative, but there are currently no toxins registered for use against stoats. We present the results of two trials undertaken in Waitutu Forest, Southland, to test the effectiveness of para-aminopropiophenone (PAPP) as a stoat control agent in the field. Meat baits containing 13 mg PAPP were placed in bait stations for five nights and tracking rates were used to monitor changes in stoat abundance. In the first trial, the index of stoat abundance was reduced by 83% and in the second trial by 87%. Our results indicate that PAPP is an effective toxin for stoats in the field and has the potential to provide a significant new tool for management of native species.

Keywords:

• control
• field trial
• para-aminopropiophenone
• PAPP
• stoat
• Mustela ermine
• toxin
• tracking tunnel

Introduction

Introduced mammalian carnivores have had a devastating impact on New Zealand's fauna (Wilson 2004). Stoats (Mustela erminea), which were introduced in the 1880s, are implicated in the decline of many of New Zealand's native birds including kiwi (Apteryx spp.) (McLennan et al. 1996), kaka (Nestor meridionalis) (Wilson et al. 1998), northern New Zealand dotterel (Charadrius obscurus aquilonius) (Dowding & Murphy 2001) and blue duck/whio (Hymenolaimus malacorhynchos) (Whitehead et al. 2008). It has been suggested that without intervention stoat predation will cause the extinction of a number of native species on mainland New Zealand (e.g. Wilson et al. 1998). The ongoing management of stoats is therefore a critical component of many New Zealand conservation strategies for protecting indigenous wildlife, especially avifauna (Dilks et al. 2003; Whitehead et al. 2008).

At present, kill-trapping is the main technique used to control stoats in New Zealand (King & Murphy 2005). However, trap lines are expensive to set up and labour intensive to service, particularly over large areas (Brown 2003). Attention is therefore turning to the possible use of poisoning as a more cost-effective means of control. Currently there are no toxins registered for control of stoats in New Zealand, although a number of compounds have been trialled. Hen eggs injected with 1080 (sodium fluoroacetate) were found to be effective (Dilks & Lawrence 2000; Spurr 2000) but were impractical in the field. Preliminary studies with cholecalciferol (Spurr et al. 2002) and diphacinone (Lawrence & Dilks 2000) were undertaken, but did not progress to the point where either compound could be registered. Stoats are known to die from secondary poisoning during control operations using 1080 and brodifacoum to target brushtail possums (Trichosurus vulpecula) and ship rats (Rattus rattus) (Murphy et al. 1998, 1999; Gillies & Pierce 1999). Stoat densities are reduced in this way, but these operations only occur in forested areas and even there, the impact on stoats may be variable, depending on the level of primary poisoning achieved.

Para-aminopropiophenone (PAPP) is now being investigated in New Zealand and Australia as a potential tool for controlling a range of introduced predators, including stoats (Marks et al. 2004; Fisher et al. 2005; Murphy et al. 2007). At appropriate doses, PAPP causes the formation of high levels of methaemoglobin, reducing the oxygen-carrying capacity of the blood and resulting in a lethal deficit of oxygen to the brain (Vanderbelt et al. 1944). Mammalian carnivores are generally more sensitive to PAPP than birds and rodents, so it has some degree of target selectivity (Savarie et al. 1983; Fisher et al. 2008; Eason et al. 2010). An effective antidote (methylene blue) is also widely available (Coleman & Coleman 1996).

PAPP has been shown to be effective against stoats in laboratory and pen trials. Stoats orally gavaged with PAPP in laboratory trials generally died within 1 hour of receiving a lethal dose (Fisher et al. 2005). In pen trials, they died within 15–85 min after eating minced rabbit (Oryctolagus cuniculus) baits containing c. 13 mg PAPP (Murphy et al. 2007). PAPP appears to be a humane toxin as its effect is relatively rapid and there were no signs of stress or vomiting associated with the poisoning (Murphy et al. 2007). The aim of this study was to determine the field efficacy of PAPP in controlling stoats, to provide data for registration purposes.

Methods

Two field trials were undertaken in a mixed Nothofagus, podocarp-hardwood forest with an open understory dominated by crown fern (Blechnum discolor) in Waitutu Forest, Fiordland National Park (46°14.4′S, 167°3.9′E). The first trial was undertaken in August 2008 at the ‘Waitutu South’ site (Fig. 1) and the second trial in November 2008 at the ‘Slaughter Burn’ site (Fig. 2).

Figure 1  Location of the Waitutu field trial at Waitutu Forest, Fiordland National Park, New Zealand. Tracking tunnel lines and PAPP bait stations are marked.

Figure 2  Location of the Slaughter Burn field trial at Waitutu Forest, Fiordland National Park, New Zealand. Tracking tunnel lines and PAPP bait stations are marked.

Stoat abundance was monitored quarterly (in February, May, August and November) with tracking tunnels. There were six tracking-tunnel lines at both the Waitutu South and Slaughter Burn sites. Each tracking tunnel line consisted of five tunnels placed at 100-m intervals. Tunnels consisted of a wooden base, a plastic removable tray and a black corrugated plastic cover. The tray was divided widthways into three equal-sized partitions. The central section contained a sponge impregnated with a 1:3 solution of red food dye diluted with water and the two outside partitions held pieces of brown paper. Tunnels were baited with fresh rabbit meat placed centrally on the sponge and left for three consecutive nights (Gillies & Williams 2007). The number of tunnels displaying stoat footprints was used to calculate an index of stoat abundance (mean percentage of tunnels tracked by stoats per line). In the first trial we compared indices of stoat abundance at the Waitutu South (treatment) and Slaughter Burn (non-treatment) sites before and after treatment to examine the effect of poison baiting (an unreplicated BACI design; Morrison et al. 2001). Changes in stoat tracking rate were calculated using the difference in tracking rates for the same lines of tunnels before and after each trial. These differences could then be averaged across lines within the treatment and non-treatment areas and the overall decrease in the mean stoat tracking rate determined. For the second trial (Slaughter Burn) we could only calculate the difference in tracking rates for the same lines of tunnels before and after each trial. Assuming unequal variances and using an arc sine transformation suitable for proportions, 95% confidence intervals were calculated.

PAPP was milled to a consistent particle size, mixed with carriers, and formed into a proprietary paste by Connovation Ltd (Connovation Ltd., PO Box 58-613, Auckland). Individual doses of the paste (c. 33 mg, containing c. 13 mg PAPP) were placed into individual sealed plastic containers for transport to the trial sites. Paste from the batch (number 4392) used in the trials was analysed by high performance liquid chromatography (HPLC) using a fluorescence detector (Landcare Research TLM056: report number T3203) to confirm the concentration of PAPP in the paste was approximately 40%.

For both trials, baits were prepared on site immediately before deployment. Baits consisted of c. 10 g of minced rabbit meat in a ball with the pre-weighed PAPP paste placed inside. A single bait was placed in each bait station and checked daily. Any bait stations recording activity (bait taken and/or tracks present) had the bait and tracking card replaced. Any uneaten baits were replaced after two days to maintain freshness. All baits were removed after the fifth night and returned to Connovation Ltd for disposal. Tracking tunnels were set for post-treatment monitoring when baits were removed.

Waitutu South field trial

The site covered approximately 900 ha east of the Waitutu River and west of the Angus Burn. Eighty-eight bait stations were positioned on an existing bird-monitoring grid at 300×300-m spacing intervals across the site (Fig. 1). Bait stations consisted of single Connovation tracking tunnels (Connovation Ltd., Auckland) containing pre-inked tracking cards so that animal visitations and bait take could be monitored. Bait-stations were not pre-baited before the trial. The Slaughter Burn tracking tunnel lines provided a non-treatment comparison for this trial.

Slaughter Burn field trial

The second trial was in the northern section of the study site (c. 600 ha), which had been used as the non-treatment site for the first trial. Ninety-nine bait stations were deployed (Fig. 2); these were a mixture of wooden kill trap tunnels but without traps (n=48) and Connovation tracking tunnels (n=51). Stations were placed 200 m apart and all were pre-baited with minced rabbit 3 weeks before the PAPP baiting. The wooden trap tunnels had mesh across the ends with a hole for stoats to enter, and most had an internal mesh ‘baffle’ with a similar-sized stoat entrance hole to limit access by other species. The tracking tunnels had no netting on the ends.

Results

Waitutu South field trial

High stoat tracking rates were recorded at both treatment and non-treatment sites for at least 18 months before the trial and 100% of the tracking tunnels were tracked at both sites immediately before baiting (Fig. 3). At the treatment site, the stoat tracking rate immediately following the trial was 17% (Fig. 3) and this increased to 27% 3 months post baiting. In contrast, the tracking rate at the non-treatment site remained high at 97% both immediately before, and 3 months after, the trial. This represents a difference in the mean stoat tracking rates between the treatment and non-treatment sites of 80% (95% CI=51.1–99.9%—arc sine transformed and assuming unequal variances).

Figure 3  Stoat tracking rates at the Waitutu South and Slaughter Burn field sites between February 2007 and November 2008. Arrows indicate deployment of PAPP at Waitutu South (August 2008) and Slaughter Burn (November 2008). Tracking tunnels were not run at the Waitutu South site post the Slaughter Burn PAPP trial (November 2008).

During the trial, 22 baits were eaten, 14 (63.6%) by stoats, seven (31.8%) by mice (Mus musculus) and one (4.6%) by a brushtail possum. In all bait stations where stoat tracks were recorded there was no trace of bait remaining, but where mice were the only visitors there were usually fragments of meat left in the tunnel. The timing of bait take showed that stoat densities were probably reduced very rapidly, with 10 of 14 (71%) baits taken by stoats in the first 2 nights (Fig. 4).

Figure 4  Timing of bait take by stoats over the five nights of the PAPP trials in the Waitutu South and Slaughter Burn field sites.

Slaughter Burn field trial

The tracking rate was 97% at the Slaughter Burn site before baiting and fell to 10% immediately following the trial (Fig. 3); a mean decrease in stoat tracking rates of 87% (95% CI=68–98%—arc sine transformed). During this trial, 19 baits were taken by stoats and none were taken by other animals. As with the first trial, most baits (79%) were taken by stoats in the first 2 nights (Fig. 4). On 3 of the 5 nights, tunnels were tracked by a stoat or stoats but the PAPP bait was not eaten—two tunnels on the second night, two tunnels on the fourth night and one tunnel on the fifth night. In the post-treatment tracking, only three tunnels were tracked and, in each case, the rabbit meat bait was not eaten (in the pre-trial tracking the bait was always missing if stoat tracks were recorded). The tunnels were in the same general area and the tracks may have been from just one stoat.

Discussion

The results reported here are from the first field trials of PAPP baits targeting stoats, and showed that baiting over 5 days reduced stoat tracking rates significantly. The fact that PAPP acts rapidly (Murphy et al. 2007) and that there was a minimum straight-line distance of 200–300 m between bait stations suggests that any individual stoat is unlikely to consume more than one bait. We believe therefore that the number of stoats poisoned will be the same as (or very similar to) the number of baits taken. Stoat densities of 2.5–4.2/km² have been recorded in South Island forest (King & Murphy 2005), so the number of stoats we assume to have been poisoned is consistent with this.

We recognise that there were methodological differences between the two trials, and that they are not strictly comparable. One was a before-after comparison at a single site, while the other was a simultaneous comparison of treatment and non-treatment sites (unreplicated BACI). There were also differences in time of year, tunnel type and layout, and pre-baiting occurred in one trial but not the other. Most of these differences resulted from logistical constraints and were unavoidable. However, in spite of their differences, the two trials produced very similar results.

Stoats are known to move over large distances relatively quickly (King & Murphy 2005) and the second trial was undertaken at a time of year when juvenile stoats are starting to disperse (King & Murphy 2005). Kill-trapping that was initiated after the PAPP trial confirmed rapid reinvasion by stoats, with eight caught in the Slaughter Burn area within 3 weeks of the trial ending (Colin Bishop pers. comm.). This highlights that a 5-night poison operation (as in this study) would only be expected to lower stoat density temporarily.

There have been few studies on the humaneness of potential stoat toxins. Stoats took between 75 and 247 min to die after eating 0.1% 1080 baits, with the period of greatest distress lasting 45 min on average (Potter et al. 2006). The average time to death of stoats feeding on 0.64% cholecalciferol bait was 6.3 days, but few signs of toxicosis were observed (Spurr et al. 2002). PAPP appears to be humane; its toxic effects are related to the rapid formation of methaemoglobin which results in hypoxia (Vanderbelt et al. 1944). It also acts rapidly—the first visible sign of toxicosis in stoats was lack of coordination 6–40 min after ingesting a PAPP bait, with death 9–45 min later and no signs of discomfort, stress, or vomiting associated with the poisoning (Murphy et al. 2007).

PAPP is readily metabolised and is unlikely to persist in carcasses at levels which will cause secondary poisoning (Savarie et al. 1983; Wood et al. 1991; Eason et al. 2010). Mammalian carnivores are also generally more sensitive to PAPP than are birds, although there is some variability in the response of different bird species (Table 1). Together, these facts suggest that PAPP is likely to cause less non-target mortality than some of the toxins currently used for pest control in New Zealand.

Table 1  Reported LD₅₀ values (the dose required to kill 50% of the sample population) for PAPP.

Species	LD50 (mg/kg)	Reference
Domestic cat (Felis catus)	5.6	Savarie et al. 1983
Dog (Canis familiaris)	7.5	Coleman et al., 1960
Stoat (Mustela erminea)	9.3	Fisher et al. 2005
Ferret (Mustela furo)	15.5	Fisher & O'Connor 2007
Red fox (Vulpes vulpes)	<25.2	Marks et al. 2004
Dama wallaby (Macropus eugenii)	89	Fisher et al. 2008
Rat (Rattus norvegicus, albino)	177	Savarie et al. 1983
Mouse (Mus musculus, albino)	223	Savarie et al. 1983
Brushtail possum (Trichosurus vulpecula)	≥500	Fisher et al. 2008
Mallard duck (Anas platyrhynchos Pekin breed)	32	Eason et al. 2010
Mallard duck (Anas platyrhynchos Pekin breed)	38	Fisher et al. 2008
Red-winged blackbird (Agelaius phoenicus)	133	Savarie et al. 1983
Blackbird (Turdus merula)	174	Eason et al. 2010
Black-billed magpie (Pica pica)	178	Savarie et al. 1983
Common crow (Corvus brachyrhynchos)	≥178	Savarie et al. 1983
Coturnix quail (Coturnix coturnix)	>316	Savarie et al. 1983
Starling (Sturnus vulgaris)	>316	Savarie et al. 1983
Weka (Gallirallus australis)	568	Eason et al. 2010
Australian magpie (Gymnorhina tibicen)	1388	Eason et al. 2010

PAPP is the first active ingredient to be developed as a vertebrate pesticide in New Zealand for 30 years, and has the potential to provide a significant new tool for management of native species. Data from these field trials, together with other required dossiers, have been submitted to the NZ Food Safety Authority (NZFSA) and the Environmental Risk Management Authority (ERMA) for the registration of a PAPP-based stoat control product in New Zealand.

Further research will include the development of additional delivery systems for PAPP. The fresh-meat baits used in our trials have a limited life, and for cost-effective control there is a requirement for either an attractive long-life bait or delivery that does not involve baiting. A system is currently being tested that delivers a dose of PAPP onto the fur of an animal passing through a tunnel. Conceptually the animal would then ingest the toxin when it groomed its fur.

Acknowledgements

Thanks to all Connovation staff directly and indirectly involved and to Duncan MacMorran, Steve Hix, Colin Bishop, Jeanette Drysdale and Grant Harper for assistance with pre-trial planning and organising the ERMA, iwi and DOC permits. The field work was assisted by Tim Grainger, Colin Bishop, Grant Harper and Alan Jones. The pre-trial tracking tunnel lines at Waitutu were run by Sean Jacques and Chris Betts. Lyn Booth and Les Brown at Landcare Research are thanked for undertaking the quality assurance analyses on the PAPP pastes used in the two trials. Thanks also to Ian Westbrooke for statistical advice and John Dowding for helpful comments on the draft manuscript.Both field trials were carried out under Approval HSC000335 from ERMA and provisional registration from the NZFSA (V9513) and were approved by the Lincoln University Animal Ethics Committee (Approval #190).