Trends in the development of mammalian pest control technology in New Zealand

ABSTRACT

Rodenticide and vertebrate pesticide registrations have declined worldwide over the last 30 years. New Zealand has not followed this trend, instead retaining essential toxins and traps, improving their use and exploring new mammal control tools. Looking to the immediate future, as well as continuing to improve the use of existing tools, there are opportunities for further advances in emerging technologies such as wireless technology for species recognition and aiding trapping programmes, self-resetting traps and toxin-delivery systems to be enhanced with advanced lures, and new toxins which increasingly combine ‘low-residue’ characteristics with selectivity and humaneness. More selective baiting and delivery systems will enable more targeted control of possums, mustelids and rodents. The use of new toxins with advantages in specific settings should be complemented by improvements in resetting trap technology, barrier approaches, and novel biocontrol and genetic concepts. Sodium fluoroacetate (1080) and other important tools have been retained; we have the ingredients for transformational change, and new tools are emerging from a research and development pipeline. However, there has been limited practical experience with emerging technologies compared with traditional or 1080 baits. Additional investment and practical experience is imperative, at this stage, to enable the potential of new toxins and other tools to reach their potential. It is also important for the future of New Zealand’s biodiversity that research continues to be focused on emerging technologies as well as on completely novel ideas.

KEYWORDS:

• Emerging technologies
• environmental safety
• investment
• mammalian pests
• native species
• New Zealand
• research pipeline
• strategy

Introduction

This review specifically focuses on research aimed at retaining the registration of proven vertebrate pesticides, on new initiatives to develop more humane and species-targeted toxins and also, briefly, on reviewing alternative control agents. It is written with special reference to pest impacts in New Zealand, but also explores international perspectives. Registration requirements for vertebrate pesticides are assessed, and the history, trends and types of toxins available are covered, as well as their advantages and disadvantages.

New Zealand wildlife evolved in the absence of mammalian predators (Parkes & Murphy 2003) so was particularly vulnerable to the consequences of their arrival. Many endemic bird species are now extinct, and the proportion of New Zealand birds classed as threatened is the highest in the world. Vertebrate pesticides are used to mitigate conservation problems caused for wildlife by the impacts of introduced mammals, such as possums (Trichosurus vulpecula), stoats (Mustela erminea), Norway rats (Rattus norvegicus), ship rats (Rattus rattus) and house mice (Mus musculus), and for indigenous plants and animals in unique ecosystems and island habitats (Dickman 1996; Innes & Barker 1999; Parkes & Murphy 2003, 2004; Towns & Broome 2003).

With invasive mammals now eradicated from many small, uninhabited New Zealand islands (Veitch et al. 2011) this country is regarded as a world leader in such eradications (Taylor & Thomas 1989, 1993; Courchamp et al. 2003; Towns & Broome 2003; Bellingham et al. 2010; Broome et al. 2010; Veitch et al. 2011; Russell & Broome 2016); however, there is a long way to go before the decline in native species across increasingly large areas of mainland New Zealand can be reversed. To address this challenge New Zealand must retain the existing toolbox, and develop new tools and strategies.

Despite the continuing need in New Zealand and worldwide for new and effective tools for the conservation and protection of agriculture from vertebrate pest impacts, over the last 50 years the number of vertebrate pesticides registered globally has declined. In the European Union (EU), the Biocidal Products Directive (98/8/EC) has substantially reduced the number of vertebrate pesticides available (Adams 2005; Buckle et al. 2005). The USA reduced the number of federal registrations from 72 in 1960 to 30 by 1998 (Ramey et al. 1992, 1994; Fagerstone & Schafer 1998). The increase in data required for new registrations overseas, and the associated financial burden, has caused industry not to continue with important but minor use registrations (Fagerstone et al. 1990). They lack a compelling ‘national good’ imperative akin to saving the iconic kiwi (Apteryx spp.) from predation as a driver for improved technology.

New Zealand has, through a combination of government and industry investment, research effort and practical management experience, retained some important toxin registrations whilst also advancing new compounds and baits with better safety profiles and improved traps. This approach has not been used by some other fields of pest management in New Zealand, to their detriment. For example, while there has been good progress in developing environmentally-friendly insect pest management options (Goldson et al. 2015), pesticides are still necessary in numerous situations for invertebrate pest control, and there are concerns that some of the older, standby, broad-spectrum pesticides are now being phased out, which could leave critical gaps in integrated pest management strategies (Goldson et al. 2015).

It is clear that the continued use of vertebrate pesticides is important, but there are ethical concerns (Hansford 2016), public criticism and market risks in the use of toxins whose registrations have been withdrawn in the USA or other markets. These are key factors in ensuring we fully understand the toxins we use can address scientific and technical questions and use them in ways that minimise risk. Issues being systematically addressed for vertebrate pest management have been the need for cost-effective, humane and targeted suppression at very large scales, and a requirement for public support for mammalian pest control or eradication (Mason & Littin 2003; Cowan & Warburton 2011), especially where this involves toxins.

There are increasing aspirations for New Zealand to move away from the use of pesticides for all pest management issues towards control systems based on good biological understanding of target species (Thresher et al. 2014; Goldson et al. 2015). A broader strategy for mammalian pest control has also explored research on biocontrol, including rabbit haemorrhagic disease (Eden et al. 2015) and non-lethal control, including more recent information from other fields of pest management such as gene editing (Thresher et al. 2014; Gantz et al. 2015). However, in the field of vertebrate pest control, failure to retain and improve existing pesticides for the next decade, as well as researching long-term alternatives, could have serious consequences. The main reasons to retain a toxin-based approach for vertebrates in the New Zealand context have been: 1. the urgent need for effective tools to protect native wildlife (Hansford 2016); 2. the successful results following their use in control and island eradication programmes; and 3. the lack of viable alternatives. Toxins and traps are the mainstay of all currently practical mammalian control programmes in urban, agricultural and conservation environments, and this will remain so for at least the next 10 to 20 years. In addition they will play important roles in enabling Predator Free New Zealand 2050 to meet its goals of dramatically reducing New Zealand’s predator populations, including rats, stoats, possums, weasels and ferrets.

In conclusion, research targeted at the management of vertebrate pests differs from that of invertebrate pests and weeds in New Zealand. There has been a deliberate plan to retain and improve the use of existing tools and expand the range of tools in the toolbox (Eason et al. 2014; Goldson et al. 2015) in parallel to research aimed at non-lethal solutions such as disrupting breeding (Tyndale-Biscoe & Hinds 2007). The focus of this review is principally on historic and recent development in toxins used in New Zealand and their future potential for controlling possums, rats and stoats. There is also a briefer commentary on other pest control technologies and strategies. Research on resetting traps and novel approaches is recent and there is, at this time, less published material to review.

History and properties of vertebrate pesticides

Some natural pesticides, such as cyanide and strychnine, have been used worldwide for hundreds, possibly thousands of years. Then, between 1940 and 1990, there was a period of prolific international research and development of new compounds. Sodium fluoroacetate (1080) was developed in the 1940s, and the first-generation anticoagulant rodenticides in the 1940s, 1950s and 1960s. Cholecalciferol and second-generation anticoagulant rodenticides were developed in the 1970s and 1980s (Eason & Wickstrom 2001).

Given the serious consequences of any failure to retain existing tools for conservation (Hansford 2016), this section focuses on describing the range of toxic compounds presently available (Table 1). More detailed reviews of the characteristics, international application and toxicology of these poisons can be found elsewhere (Prakash 1988; Hayes & Laws 1991; Buckle & Eason 2015). Repellents, which can be very important in reducing non-target risk, are not reviewed here but are covered in recent publications (Cowan et al. 2016; Crowell et al. 2016a, 2016b).

Table 1. Classes of compounds or approaches used to control terrestrial vertebrate pests (adapted from Eason et al. 2010a). Note focus of the review is on mammals. Other target species and toxins included for interest or cross referencing purposes only.

Target action	Class or approach	Target organism	Active ingredient
Lethal control agents	Anticoagulant toxicants	Mammalian	First-generation  Pindone  Warfarin  Coumatetralyl  Diphacinone  Chlorophacinone Second-generation  Brodifacoum  Difethiolone  Flocumafen  Difenacoum  Bromadiolone
	Acute-acting toxicants	Mammalian	Alpha chloralose Zinc phosphide* Cholecalciferol Bromethalin Strychnine Sodium cyanide Sodium fluoroacetate (Compound 1080) Thallium Para-aminopropiophenone (PAPP)* Red squill (scilloroside) Sodium nitrite*
		Avian	Alpha chloralose 3-chloro-p-toluidine HCL (DRC-1339)
		Reptilian	Acetaminophen (paracetamol)
	Fumigants	Mammalian	Aluminium phosphide Magnesium phosphide Sodium nitrate, carbon (gas cartridge)
		Reptilian	Methyl bromide
Non-lethal control agents	Repellents	Mammalian	Thiram Egg-acrylic Predator odours Dried blood based products
		Avian	Pulegone Methyl anthranilate Anthraquinone 4-aminopyridine Methiocarb Cinnamamide Caffeine Garlic oil
	Contraceptives and novel genetic approaches	Mammalian	Gonadotropin releasing hormone* Porcine zona pellucida* Trojan females and other genetic approaches†
		Avian	Nicarbazin

*Recently registered compounds in New Zealand.

^†Effective in invertebrates under investigation in other species (Thresher et al. 2014).

Acute-acting compounds

Prior to 1950 all vertebrate pesticides were acute-acting toxicants or fumigants. After the introduction of warfarin and the other anticoagulants, the importance of these non-anticoagulants was reduced, at least for rodent control. Following the emergence of physiological resistance in some populations of rodents, the discovery of residues of the second-generation anticoagulants in wildlife (Young & De Lai 1997; Stone et al. 1999; US EPA 2008) and questions about the humaneness of second-generation anticoagulants in larger vertebrate pests (Littin et al. 2002; Mason & Littin 2003), interest in non-anticoagulants (or at least less-persistent ‘low residue’ pesticides) has revived, and new acute substances investigated (Eason et al. 2013, 2014; Shapiro et al. 2016a, 2016b, 2016c, in press). Hence the long list of known compounds in Table 1.

The principal acute vertebrate pesticides used in New Zealand are 1080, cyanide and cholecalciferol. Recent registrations include zinc phosphide, para amino-propiophenone (PAPP) and sodium nitrite (Eason et al. 2013, 2014; Shapiro et al. 2016a, 2016b, 2016c, in press). Recent registrations and developments, and the history, properties and applications of the available compounds are outlined briefly below.

Sodium fluoroacetate

Sodium fluoroacetate (1080) was first synthesised in Belgium in 1896, but was not seriously investigated as a pesticide until the 1940s when wartime shortages of strychnine and red squill stimulated the development of other toxicants (Atzert 1971). Fluoroacetate can be found naturally at lethal concentrations in poisonous plants (de Moraes-Moreau et al. 1995; Twigg et al. 1996a, 1996b) and, in the USA, 1080 is used solely for localised and very target-specific predator control in the Livestock Protection Collar (LPC), mainly to protect sheep (Ovis aries) against coyotes. In Australia and New Zealand it is formulated into baits to kill a range of introduced mammalian pests. New Zealand is the largest user of 1080 in the world, employing the active ingredient at the rate of c. 1.0 to 3.5 tonnes per year (Innes & Barker 1999), but that is less than 0.1% of the total pesticide active ingredient used annually in New Zealand (Eason et al. 2010a).

1080 is very poisonous with an LD50 of ≤ 1 mg/kg for many species (Atzert 1971; Eisler 1995). The mode of action of fluoroacetate is by inhibition of energy production in the tricarboxylic acid (Krebs) cycle, resulting in death from heart or respiratory failure (Egeheze & Oehme 1979; Eason et al. 2011). In mammals, symptoms of poisoning appear between 0.5 and 3 hours after ingestion, and most animals receiving a lethal dose die within 24 hours. All traces of the toxin are likely to be eliminated within a week.

There has been some debate about animal welfare and 1080 use (Sherley 2007). Whilst it is not as humane as PAPP (Eason et al. 2014) or cyanide (Eason et al. 2008), it is considered more humane than many others, including strychnine or anticoagulant poisons (Eason et al. 2011). The use of 1080 for pest management attracts opposition from sectors of society (Ogilvie et al. 2010; Hansford 2016), so even after 65 years of research and practical experience, there is continuing demand for research to increase its safety (Crowell et al. 2016a, 2016b), and calls for monitoring and investigation of alternative methods. In 1994, the Royal Society of New Zealand held a science workshop on 1080, which brought together international delegates with backgrounds in wildlife management, animal welfare, and biochemistry and ecology. The proceedings remain relevant as a comprehensive overview of the science underpinning the use of 1080 (Seawright & Eason 1994). In the 23 years since then, 1080-related research and monitoring has investigated bait quality and greatly reduced sowing rates, potential sublethal effects, animal welfare issues and secondary poisoning risk, reducing risk to non-target species, ecotoxicology and fate in water, soil, plants and animals (Eason et al. 2011; Northcott et al. 2014; Crowell et al. 2016a, 2016b).

Zinc phosphide

Zinc phosphide was first used as a rodenticide in 1911 in Italy (Marsh 1987). It is an effective acute field rodenticide, and was the most widely used rodenticide worldwide until the introduction of anticoagulant compounds in the 1940s and 1950s. It is currently registered in the EU, but has limited use compared with anticoagulants. It is still used as a rodenticide in the USA, Australia, the Asia-Pacific region and China (Marsh 1987; US EPA 1998; Eason et al. 2012a). It remains the toxin of choice for field use in some situations, for example against mouse plagues in Australia, and can be rapidly broadcast from ground spreaders or aircraft (Brown et al. 2002). The reason it has found favour in the USA and Australia is its lack of persistence and comparatively low risk of secondary poisoning following its field use (because it breaks down in contact with moisture) when compared with strychnine or 1080. Nevertheless, considerable care must be taken when using zinc phosphide because it still has the potential (like other toxins) to cause primary poisoning of non-target species, and treatment of accidental poisoning is difficult.

Zinc phosphide paste was approved for use as a possum control agent in New Zealand by the Environmental Protection Authority (EPA) in August 2011. Microencapsulated zinc phosphide (MZP) is acute-acting when delivered at toxic doses in baits to possums, with clinical signs first appearing from 15 minutes, and death after a lethal dose generally occurring in 3–5 hours. Death is mediated by a combination of cardiac failure and respiratory failure (Osweiler et al. 1985; Prakash 1988; Hayes & Laws 1991), which is considered moderately humane (Fisher et al. 2004). The emetic action of the zinc portion reduces the toxicity of zinc phosphide to some non-target species, but that does not protect rats, which lack a vomiting reflex (Marsh 1987). To combat that effect, a microencapsulated form of zinc phosphide has been developed for use in paste, and may be developed for solid cereal baits, initially for controlling possums and as a rodenticide (Shapiro et al. 2016c).

Despite extensive use overseas there has been only limited research and practical experience with zinc phosphide in New Zealand, especially when compared to 1080-containing baits. Additional research efforts and practical experience could enable the effective use of zinc phosphide as a conservation tool for bovine tuberculosis (TB) vector control in New Zealand (Eason et al. 2012a; Shapiro et al. 2016a). In the future, zinc phosphide could be delivered in a resetting toxin-delivery system (Blackie et al. 2016) and is being explored for rabbit (Oryctolagus cuniculus) control (J. Ross, Lincoln University, pers. comm. 2016), in part because of secondary poisoning concerns associated with aerial application of pindone baits for rabbit control and the potential public opposition to aerial 1080 rabbit control.

Cholecalciferol

Cholecalciferol (vitamin D3) is synthesised in animal skin by the action of sunlight on its precursor, 7-dehydrocholesterol. Natural dietary sources of vitamin D include fish, liver, fish oils, egg yolk, and milk fat and plants, and background levels are detectable in the blood and tissues of all mammals (Fairweather et al. 2013). To become biologically and toxicologically active, cholecalciferol must be converted to 25-hydroxycholecalciferol (25OHD). The latter metabolite is the most biologically active form of vitamin D3, which can cause calcification and death from heart failure (Dorman & Beasley 1989). Metabolism studies have shown that blood concentration of 25-hydroxycholecalciferol increases in animals receiving a near-lethal dose, reaching a peak after approximately 5 days. Elevated tissue concentrations may persist in survivors for several weeks. Time to death after a lethal dose is similar to that for rodents exposed to lethal doses of anticoagulants, and is usually 3–7 days (Marshall 1984).

Cholecalciferol was developed as a rodenticide in the 1980s (Marshall 1984; Tobin et al. 1993). The single-dose LD50 for cholecalciferol in Norway rats (Rattus norvegicus) and house mice (Mus musculus) is very similar (c. 40 mg/kg), but there is considerable species variation in susceptibility among other mammals, and possums are more susceptible.

In New Zealand cholecalciferol was first registered in the 1990s in baits at 0.4% and 0.8%, and in the USA at 0.1%. In the EU it was registered at 0.1% until this registration was discontinued, but re-registration is under consideration (R. Sharples, BASF, pers. comm. 2013), and it is not registered in Australia. Its registration in New Zealand has been extended to include both possums and rats, and it is a more affordable but still effective bait containing 0.4% cholecalciferol (Hix et al. 2012).

The most distinguishing features of cholecalciferol are its low risk of secondary poisoning of dogs (Canis familiaris) and low toxicity to birds (Eason et al. 2000). Non-target acute toxicity studies have been conducted on mallard ducks (Anas platyrhynchos), chickens (Gallus gallus domesticus) and canaries (Serinus spp.). Following oral gavage of cholecalciferol concentrate at 2000 mg/kg, there were no adverse effects in ducks. Chickens and canaries were more sensitive, and some deaths followed dosing at 2000 mg/kg.

However, primary poisoning of non-target wildlife or pets is a risk if they have access to cholecalciferol-containing bait. Feeding poisoned animals to cats, dogs, ferrets (Mustela putorius furo) and turtles (Testudines spp.) indicates that the risk of secondary poisoning is low, despite the likely presence of elevated concentrations of 25-hydroxycholecalciferol in carcasses. Low doses of cholecalciferol have been added to anticoagulant-containing baits to increase their effectiveness. Combining aspirin with cholecalciferol has been shown in recent research to greatly improve the cost-effectiveness and animal welfare impact of cholecalciferol alone (Morgan et al. 2013).

Cyanide

Cyanide disrupts energy metabolism by preventing the use of oxygen in the production of energy, causing cytotoxic hypoxia in the presence of normal haemoglobin oxygenation. When the dose is optimised, the cytotoxic hypoxia depresses the central nervous system, the most sensitive site of anoxia, resulting in rapid respiratory arrest and death (Osweiler et al. 1985; Gregory et al. 1998). Hence, of all the poisons currently used for possum control, cyanide when delivered in an optimised delivery system is considered the most humane (Gregory et al. 1998), because of the short time to death of 15–20 minutes and short duration of symptoms of poisoning. When developing new toxins for other vertebrate pests, researchers are attempting to attain the standard achieved by cyanide in possums, and avoid compounds with more protracted effects, such as brodifacoum (Littin et al. 2002).

Cyanide, as a vertebrate pesticide, is predominantly used to control coyotes (Canis latrans) in the USA and it is delivered using the M-44 mechanical ejector (Blom & Connolly 2003; Fagerstone et al. 2004). When the top is pulled by a predator, the device ejects cyanide into the mouth of the animal. Since its registration in 1997, cyanide has become an accepted method for ground baiting for possum and wallaby (Macropus spp.) control in New Zealand (Eason et al. 2010d; Ross et al. 2011; Shapiro et al. 2011). In Australia it has been used experimentally for killing foxes (Marks & Gigliotti 1996; Marks et al. 2004). The introduction of an encapsulated cyanide pellet in the 1990s, which kills possums without secondary poisoning, underpinned the extension in 2009 of the encapsulated cyanide pellet registration to include wallabies (Eason et al. 2010c; Ross et al. 2011; Shapiro et al. 2011).

Para-aminopropiophenone

PAPP was originally studied as a treatment for cyanide and radiation poisoning in the 1940s (Rose et al. 1947; Eason et al. 2014). It is toxic to carnivores, although birds and humans are less sensitive (Savarie et al. 1983; Fisher & O’Connor 2007; Murphy et al. 2007; Eason et al. 2010a, 2010b). This is primarily because eutherian carnivores have metabolic pathways different from those in other orders of animals (Wood et al. 1991). The registration of PAPP in New Zealand builds on initiatives at the National Wildlife Research Center (NWRC) in the USA (Savarie et al. 1983), toxicology research in the UK on methaemoglobinaemia inducers (Marrs et al. 1991), research in Australia (Marks et al. 2004; Cowled et al. 2008) and by the New Zealand Department of Conservation (DOC) (Murphy et al. 2007).

The toxic effects of PAPP are related to its ability to reduce the oxygen carrying capacity of the red blood cell through the formation of methaemoglobin. The onset of symptoms is rapid, and cats and foxes are usually unconscious within 30–45 minutes (Marks et al. 2004). This leads rapidly to a lack of oxygen to the brain and other vital organs, and death due to respiratory failure. Normally, methaemoglobin concentration in the blood is below 1%. Levels of methaemoglobin in the blood above 70% are usually fatal, creating a lethal deficit of oxygen in cardiac muscle and the brain. In stoats and feral cats, clinical signs after a lethal dose first appear in 10–20 minutes for stoats and at around 35 minutes for cats; death usually follows within 2 hours. Affected animals become lethargic and sleepy before they die, hence PAPP is classed as relatively humane. Methylene blue will reverse the methaemoglobinaemia induced by PAPP, so is considered a viable antidote to PAPP exposure (Rose et al. 1947).

PAPP has been developed in New Zealand specifically for the control of stoats and feral cats (Felix catus) because of the special sensitivity displayed by these species (Eason et al. 2014). PAPP paste was approved as a stoat control agent in New Zealand by the EPA in August 2011, and for feral cat control in November 2011 (Eason et al. 2010b, 2010c; Eason et al. 2014). Registration of PAPP in Australia for control of foxes and feral cats was achieved in 2016 (Fleming et al. 2006). Over the last 10 years, New Zealand-based research has focused on determining its toxicity to predators, field effectiveness for controlling stoats and feral cats, animal welfare profile, toxicology, ecotoxicology, and understanding and reducing non-target risks.

When meat baits containing toxic doses of PAPP are applied in bait stations in field settings, stoat and feral cat numbers can be rapidly reduced (Dilks et al. 2011). However, there has been limited practical experience with PAPP to date, especially when compared with traditional traps or 1080 baits. Additional practical field experience is essential, in different delivery vehicles, to enable the effective use of PAPP to help protect native species from introduced predators. In the future, PAPP may be developed in bait suitable for both ground and aerial control (E. Murphy and M. Crowell, DOC, pers. comm. 2016) and in a resetting toxin-delivery system (Murphy et al. 2014a).

Sodium nitrite

Sodium nitrite, a meat preservative, was first investigated as toxin for feral pigs (Sus scrofa) in Australia (Cowled et al. 2008). Of all the species tested, pigs are among the most sensitive to the chemical on a mg/kg basis: time to death for feral pigs is 2–3 hours with few visual symptoms (Cowled et al. 2008). In December 2013, following 5 years of research into its toxicity to possums and feral pigs, field effectiveness, animal welfare profile, toxicology, ecotoxicology, understanding secondary poisoning and reducing non-target risks (Shapiro et al. 2016a, 2016b, in press), encapsulated sodium nitrite (ESN) was registered by the New Zealand EPA for the control of both possums and feral pigs.

This new toxin represents an advance in terms of humaneness, safety and non-target impacts. It is a methaemoglobinaemia inducer, with a mode of action on the red blood cells similar to that of PAPP. Both rank consistently high on criteria of humaneness and the ready availability of an antidote (methylene blue). The hearing for public consultation required before EPA approval garnered support from TBFree NZ, Auckland Regional Council, DOC and members of the public (including the hunting fraternity) concerned about secondary poisoning of dogs, because that is most unlikely following use of sodium nitrite baits. Sodium nitrite is also the first (and currently only) toxin available for feral pig control in New Zealand. Pen and field trials with feral pigs have demonstrated high levels of efficacy and welfare when the toxin is applied at optimum doses in baits (Shapiro et al. 2016a, 2016b, in press).

Alphachloralose

Alphachloralose is a narcotic with a rapid effect. It slows a number of essential metabolic processes, including brain activity, heart rate and respiration, inducing hypothermia and eventual death. It is most effective against small rodents such as mice in cold or cool conditions. In the UK, alphachloralose is most often used in baits containing 2%–4% of the active material for mouse control (Buckle & Eason 2015).

In a number of countries there is some use of this compound for controlling bird pests; clearly, because of its toxicity to birds, it must be used with care when applied in baits for control of mice. In New Zealand it is registered as an avicide.

Norbormide

Norbormide is a selective rat toxicant. It was developed in the 1960s, but its use was discontinued in the 1970s as anticoagulant toxins became more popular. Taste aversion has in the past limited its effectiveness, and field efficacy results were generally poor. After a lethal dose most animals die within 8–12 hours. As it is comparatively acute-acting it is likely to be more humane than most other rodenticides, because of the relatively short time to death and duration of symptoms of poisoning when compared with anticoagulant rodenticides and cholecalciferol.

Norbormide causes vasoconstriction (narrowing) of small arteries and vasodilation (widening) of large arteries in rats (Roszkowski 1965), which causes a rapid fall in blood pressure. Death probably results from circulatory disorders and heart failure (Cavalli et al. 2004; Ricchelli et al. 2005).

Norbormide is highly toxic to members of the genus Rattus compared with other mammals or birds (Roszkowski 1965). Rats are 150-fold and 40-fold more sensitive to norbormide than mice and guinea pigs (Cavia porcellus), respectively, while most other mammals and birds tested are > 100-fold less sensitive (Roszkowski 1965). Prodrug forms of norbormide have been developed that aim to delay the action of the toxicant and increase palatability by masking the taste (Rennison et al. 2012, 2013). Two New Zealand research teams, one currently funded by the DOC Predator Free 2050 fund (http://www.landcareresearch.co.nz/publications/newsletters/kararehe-kino/kararehe-kino-issue-29/host-selective-toxins) supporting Landcare Research with Orillion (formerly ACP Ltd) and the second through investment by Connovation Ltd, are looking at different ways of improving the effectiveness of norbormide and producing it in forms which are more palatable (D. MacMorran, Connovation Ltd, pers. comm. 2017; Jay-Smith et al. 2016). Connovation Ltd has a formulation that is equally effective in Norway and ship rats, and further pen and field trials are planned by both groups in 2017–2018. If either or both of these groups are successful it will be a huge advance for targeted pest control with no non-target impact.

Challenges ahead include: sufficiently delaying the action of the toxicant in rats, effectively preventing the appearance of symptoms of poisoning until a lethal dose has been ingested; maintaining high toxicity; and incorporating the norbormide formulation into a bait matrix to produce a product that would be sufficiently attractive to all rats in the wild to deliver a lethal dose while being relatively economic. At the current rate of development it is expected new forms of norbormide could be registered and available for field use within the next 5 years (Campbell et al. 2014). Because the action of norbormide is unique, and linked to rodent-specific receptors, genome screening for pest-specific toxin receptor targets could produce new toxins for other species in the future (B. Hopkins, Landcare Research, pers. comm. 2015).

Bromethalin

Bromethalin was developed in the 1970s. It is a single-feed rodenticide that is registered for use only in the USA, and restricted to bait stations in and around buildings for the control of commensal rodents (Dreikorn 1978). In the 1970s bromethalin was considered for use in New Zealand, but regarded as inhumane, so it is not registered here or in the EU. Bromethalin is a neurotoxicant. It has no special advantages and, unlike cholecalciferol, it is toxic to birds (Buckle & Eason 2015).

Strychnine

Strychnine is found in the seeds of the tree Strychnos nux-vomica. It is an alkaloid used for rodent and vertebrate pest control since the mid-1800s (Schwartze 1922). It is registered only in the USA for underground use against some rodent species such as pocket gophers (Geomyidae spp.) and moles (Talpidae spp.), but not currently in New Zealand or the EU.

Strychnine is an acute-acting compound, and poisoned animals often die in less than 1 hour as a result of respiratory failure (asphyxia), occasionally in 24 hours or longer. The typical signs of strychnine poisoning are restlessness and muscular twitching, which progress to convulsive seizures and violent muscular spasms before death (Osweiler et al. 1985). Strychnine was used in New Zealand but has been deliberately phased out because of its inhumaneness and persistence (Eason & Wickstrom 2001). Inhumaneness is linked to the very pronounced and sometimes prolonged unpleasant symptoms of poisoning when animals are still conscious.

Red squill

Red squill was extracted from bulbs of the Mediterranean plant Urginea maritima. The active substance is the cardiac and nerve glycoside, scilliroside (Hone & Mulligan 1982; Meehan 1984). The compound was considered to be as effective against Norway rats as other acute toxins, but its efficacy against ship rats and house mice is questionable (Meehan 1984).

Symptoms of poisoning include hind limb paralysis, convulsions, emesis (except in rodents) and diarrhoea. Atropine sulphate may be antidotal. Red squill is banned in the UK under the Animal (Cruel Poisons) Act 1962, particularly because of its properties as a convulsant, and has not been pursued in New Zealand.

Thallium sulphate

Thallium sulphate achieved a considerable degree of popularity as a rodenticide because it is readily accepted in baits and highly toxic to all rodent species (Mallis 1960). But, like a number of other acute-acting toxins, it suffers from the disadvantages of high toxicity to non-target animals, and the lack of any antidote. Its prohibition in most countries is no loss because it has no advantages over other acute toxins currently registered in New Zealand.

First-generation anticoagulants

First-generation and second-generation anticoagulant rodenticides (Table 1) have the same mode of action, interfering with the synthesis of blood clotting factors resulting in haemorrhaging and death. Over 10 anticoagulant agents have been synthesised, and their principal use worldwide in pest control has been against commensal rodents, primarily Norway rats, ship rats and house mice. Many of the anticoagulants listed in Table 1 are registered for commensal rodent control in the USA, Australia and New Zealand; in all three countries they have various field use applications in conservation programmes.

Warfarin

Warfarin was one of the earliest first-generation anticoagulant rodenticides. It has been used in a range of rodent baits since it was first introduced in 1947. Warfarin was developed following the discovery that coumarins were responsible for haemorrhaging in cattle eating mouldy clover hay in the USA in the 1930s. Like the other anticoagulants it inhibits the synthesis of vitamin K-dependent clotting factors. In addition warfarin is reported to induce capillary damage.

In general the symptoms of poisoning do not appear suddenly, and in rats will culminate in death within 5–7 days of the first ingestion of a lethal dose. The single dose LD50 is 50–100 mg/kg in rats (species unspecified), or daily doses of 1 mg/kg for 5 days (Osweiler et al. 1985), which will kill rats in 5–8 days. Warfarin was used for the control of feral pigs in Australia (Choquenot et al. 1990), but this use has been phased out.

Pindone

Pindone belongs to the indandione class of anticoagulants, which differ chemically from coumarin anticoagulants such as brodifacoum or warfarin. It was synthesised in 1937 and developed as a pesticide in the early 1940s. Pindone has been used to control rodents and even possums in New Zealand, but to a lesser extent since the introduction of more potent second-generation anticoagulants such as brodifacoum. However, it remains favoured and effective for rabbit control in Australia and New Zealand (Eason & Jolly 1993), and a single dose of approximately 18 mg/kg is sufficient to kill a rabbit. In rabbits the repeat dose (7 days) LD50 is 0.52 mg/kg/day (Hone & Mulligan 1982).

Refinements in the use of pindone baits, and better strategies for rabbit control operations, are still important. Recent research has emphasised the risks of secondary poisoning (Fisher et al. 2015), although this is less likely than if a second-generation anticoagulant was used in a similar manner. Pindone acts like the other anticoagulant toxicants by interfering with the normal synthesis of vitamin K-dependent clotting factors in the liver. The weaker potency of first-generation anticoagulants such as pindone is related to its generally lower binding affinity compared with second-generation compounds (Parmar et al. 1987; Huckle et al. 1988). As with all other anticoagulant compounds, clinical signs of toxicosis in animals will usually reflect some manifestation of haemorrhage (Osweiler et al. 1985).

Diphacinone

Diphacinone is another first-generation anticoagulant of the indandione class. Diphacinone is more toxic than are warfarin and pindone to most rats and mice (Buckle & Smith 2015). In New Zealand it is registered in cereal paste bait primarily for the field control of rodents, and it has also been incorporated into fish-based bait for ferrets. In the USA it has been registered by the US EPA for field use against rodents for conservation purposes, providing an alternative to brodifacoum (Eason et al. 2010a).

Diphacinone, like other anticoagulants, inhibits the formation of vitamin K-dependent blood clotting factors. Clinical and post-mortem signs of toxicosis are as for other anticoagulants. The persistence of diphacinone in the liver is similar to that of pindone, and both are rapidly eliminated so do not bioaccumulate as do the second-generation anticoagulants (Fisher et al. 2003). Chlorophacinone has properties similar to diphacinone, but with slightly greater potency. It is registered in the USA, but not in New Zealand (Eason et al. 2010a) and hence is not discussed further in this review.

Coumatetralyl

Coumatetralyl was developed in 1957, and is more potent than warfarin or pindone. It is used as a tracking powder (it clings to the fur, to increase uptake through grooming) or as a cereal bait, wax block or paste for rodent control. Like other anticoagulant rodenticides, coumatetralyl inhibits the formation of vitamin K-dependent blood clotting factors. It is less persistent (in sublethally poisoned animals) than is brodifacoum, but is more persistent than diphacinone (Parmar et al. 1987; Fisher et al. 2003).

Second-generation anticoagulants

The second-generation anticoagulants, such as brodifacoum and bromadiolone, are more toxic than first-generation anticoagulant rodenticides (Eason & Wickstrom 2001). They also have an unusual persistence compared with first-generation anticoagulants (Fisher et al. 2003, 2004; Crowell et al. 2013) because they are not fully metabolised and excreted before death. Their superior potency is related to their greater affinity for vitamin K-epoxide reductase, and their associated greater potential to affect non-target wildlife compared to first-generation anticoagulants is a consequence of their subsequent accumulation and persistence in the liver and kidneys after absorption (Parmar et al. 1987; Huckle et al. 1988). Only three examples from Table 1 will be discussed since they all have similar properties.

Brodifacoum

The rodenticidal properties of brodifacoum were first described in the early 1970s (Hadler & Shadbolt 1975). Brodifacoum differs from the first-generation anticoagulants in that it is very potent and can induce death in a number of animal species after only a single dose. It has been used successfully in recent rodent eradication programmes on offshore islands to protect populations of endangered indigenous birds (Taylor & Thomas 1989, 1993; Courchamp et al. 2003; Towns & Broome 2003), and to control possums and rodents in mainland New Zealand.

Second-generation anticoagulants have an important role in combating resistance to first- generation anticoagulants. However, in New Zealand their field use has been controversial because they are known to cause wildlife contamination (Young & De Lai 1997; Stone et al. 1999; Eason et al. 2002a). Wildlife contamination can extend to native birds as well as game species where there is field use of second-generation anticoagulants (Young & De Lai 1997; Eason et al. 2002). To reduce wildlife exposures and ecological risks, the US EPA has phased in restrictions in regards to the field use of second-generation anticoagulant products. Except for use around livestock facilities, baits can be applied only by professional operators, and applications must be made no further than 50 feet away from any building (US EPA 2008).

In New Zealand the problems associated with persistence of brodifacoum have been compounded by concerns about its inhumane effects on larger vertebrate pests such as possums (Littin et al. 2004). The DOC does not aerially distribute baits containing brodifacoum for routine mammal control on the mainland, restricting this kind of use to ‘one-off’ eradication operations on offshore islands or discrete use in other settings. On the other hand, brodifacoum has made a huge contribution to pest control and conservation, and is an effective and extremely important pest control tool pivotal to island pest eradication programmes around New Zealand and worldwide (Veitch et al. 2011).

Flocoumafen

Flocoumafen and brodifacoum are very similar in terms of their chemistry, biological activity and potency, persistence and risk of secondary poisoning. Flocoumafen is a second-generation anticoagulant that was developed in the early 1980s (Huckle et al. 1988). Flocoumafen has been used against a wide range of rodent pests including the principal commensal species. It is also effective against rodents that have become resistant to other anticoagulant rodenticides.

Bromadiolone

Bromadiolone also has chemical and biological effects similar to those of brodifacoum, but is slightly less potent, and it has encountered resistance in rodents after repeated use overseas. Like brodifacoum it was developed in the 1970s. Bromadiolone interferes with the vitamin K1-dependent blood clotting factors when a lethal or sublethal dose is ingested. Bromadiolone is effective against rodents that have become physiologically resistant to first-generation anticoagulant rodenticides (Eason & Wickstrom 2001).

In conclusion to this section, new acute-acting toxins have been researched and developed in New Zealand, and we continue to use anticoagulants effectively and with caution. Research programmes have advanced new toxins, baits and baiting strategies to improve humaneness and reduce the risk for non-target impacts. There is a good understanding of the properties, advantages and disadvantages of a number of non-anticoagulant and anticoagulant toxins, and this knowledge is an important contribution to their safe and effective use.

Approval and assessment

Before any new toxins or new products can be submitted for approval and registration, fundamental and applied research on animals in the field or in vitro must be done to test the risk and benefits of possible new tools. In addition, the same tests have been conducted on older compounds, such as 1080, in anticipation of a re-assessment process (Eason & Turck 2002). The principles that have underpinned the assessment of risk influence the design of the test programmes because they must address safety and efficacy questions through a logical process (Eason et al. 2010a). Research of this nature investigates the potential of a chemical to cause genetic mutations, foetal abnormalities and target-organ toxicity in humans, and toxicity to non-target species.

Toxicology studies are complemented by field trials relevant to the final use of the bait or device. Understanding the likely exposure risk of non-target species as determined by well-designed field trials will be as, if not more, important than completion of guideline laboratory studies defining hazards (Eason et al. 2010a, 2013).

In New Zealand the requirements of the Hazardous Substances and New Organism Act 1996 (HSNO) and the Agricultural Chemistry and Veterinary Medicines Act 1997 (ACVM) must be met, meaning that approvals are required from both the EPA and the Ministry of Primary Industries (MPI); consultation with Māori is a prerequisite. Welfare considerations are a key component of the ACVM registration assessment process for vertebrate pesticides, and play an important role in the selection of tools and the way they are used.

These requirements have stimulated the development of wide ranging research programmes, and the development of detailed databases for key vertebrate pesticides listing their chemistry, residues, efficacy and non-target species susceptibility, as well as ecotoxicity, toxicology, metabolism and pharmacokinetic studies. For example, in New Zealand, probably more than NZ$20 million was spent by a consortium of stakeholders over a 15–20-year period on research, consultation with community groups and updating research-based registration dossiers (Table 2) for the re-assessment process of 1080 that was completed in ERMA (2007), described in detail below.

Table 2. Studies that provided data for modern-day standards enabling the re-registration of 1080.

Study types	Status of 1080 database
Acute toxicity	Extensive literature (see Atzert 1971; Seawright & Eason 1994; Eisler 1995)
Skin and eye irritation	Completed (see Fagerstone et al. 1994)
Skin sensitisation	Completed (see Fagerstone et al. 1994)
Mutagenicity studies	Completed (see Eason et al. 1999)
Three-month feeding studies	Completed (see Eason & Turck 2002b)
Developmental toxicity (teratogenicity, reproductive toxicity)	Completed (see Eason et al. 1999)
Metabolism/pharmacology studies	Extensive published database

Such a great effort was appropriate because New Zealand is the largest user of 1080 in baits sown aerially for possum and rodent control to reduce damage to native forest and eradicate bovine TB. Likewise, a similar re-registration review has been undertaken in Australia by the Australian Pesticide and Veterinary Medicine Authority (APVMA 2008). In the USA, the NWRC completed about 250 studies costing US$3 million for resubmission of existing suitable data on 1080, zinc phosphide and cyanide (Fagerstone et al. 1990; Ramey et al. 1994; Fagerstone & Schafer 1998). Cholecalciferol has the advantage of a low risk of secondary poisoning to mammals and low risk to birds, yet its registrations have recently been discontinued in the EU (Eason et al. 2000). The data requirements of the EU biocides Directive were deemed excessive, and extremely costly to generate relative to their scientific merit and sales volume by the registrants (Adams 2005; Buckle et al. 2005; Eason et al. 2010a). This means that its utility as a tool to control rodents resistant to anticoagulants in the EU has potentially been lost.

In New Zealand, a specific risk versus benefit analysis is taken into account in the EPA approval process. For example, introduced stoats are devastating native flightless birds, including the iconic kiwi (Apteryx spp.) whose numbers have been falling throughout the country (Innes & Barker 1999). Killing stoats with poisoned baits containing PAPP in bait stations in remote New Zealand forest ecosystems obviously presents different risk/benefit scenarios from spraying insecticides and herbicides on food crops. Stoats in the environment will present a real risk to birds, whilst the risk to the same birds from PAPP baits for predator control will be minimal in comparison to predation risk.

Appropriate formulation is a key part of risk mitigation, as vertebrate pesticide products can be designed to be safe to handle and to be attractive to target individual pest animals and repel non-target species (Cowled et al. 2006; Cowan et al. 2016; Crowell et al. 2016a, 2016b). Delivery systems that facilitate contact with target species and minimise non-target species exposure are also critical (Campbell et al. 2014; Cowan et al. 2016).

The level of research required is illustrated by the new predacide PAPP, which required development costs of more than NZ$6 million before registration in both New Zealand and Australia (Eason et al. 2014). Gaining practical experience with new tools can also be challenging. For example, after a decade of research, PAPP was approved for use to control stoats and feral cats, but with label restrictions that prevented it being used effectively. The approval granted in March 2011 imposed requirements to notify all landowners and occupiers within a 3 km radius before beginning a control operation. This requirement was intended to protect domestic cats. Upon further analysis it was found that this control made it impractical to use PAPP for the control of stoats. Four years later, in July 2015, the registrant, aided by the DOC, completed negotiations with the EPA to lift this requirement to enable ground control operations to go ahead.

In conclusion to this section, researchers must be aware of EPA and MPI requirements, and Māori considerations. The data requirements for emerging technologies, as well as for completely novel genetic approaches, are not always fully defined. Experience has shown that early consultation is imperative for effective research and development programmes.

Pipeline thinking: retaining and improving existing and new technologies

Disruptive technology can displace current technology in ways that the market does not expect, but in mammalian pest control we need this effect, along with step-wise improvements that can be achieved in the short and medium term (Table 3). Importantly, collaboration across disciplines, hands-on experience of research in science or engineering disciplines relevant to product development and commercialisation, and practical wildlife management experience are the critical capabilities needed to deliver and ensure the continued improvements (Allen et al. 2014).

Table 3. Summary of recent toxin and resetting toxin delivery research milestones and tools which require additional research to optimise their use.

Type	Pre 2010	2010–2016
Products that contain vertebrate pesticides already in use in New Zealand	Research underpins 1080 being re-registered and used more safely Encapsulated cyanide pellet for possums and wallabies Low dose (0.4%) cholecalciferol Deer repellents	New approaches to 1080 aerial application Continued development of deer and bird repellents to minimise non-target impacts
Products that contain vertebrate pesticides NOT already in use in New Zealand		Zinc phosphide registered for possums Resetting device for possums tested in the field with proof-of-concept (POC) with zinc phosphide and food lures Species recognition technologies with remote sensing linked to toxin delivery D+C researched and registration filed with EPA for control of possums Progress on norbormide
New Zealand advances and registrations of a new generation of vertebrate pesticides		PAPP researched and registered for control for stoats and feral cats Resetting device for stoats tested in the field (POC) with exciting avenues for stoat lures Resetting device developed for rats and cats (POC). PAPP for ferrets (POC) Sodium nitrite researched and registered for possums and feral pigs

Research approaches needed for retaining and developing new technologies can be viewed within two frameworks, which we describe as: 1. a three-pronged approach; and 2. a pipeline of innovation (Table 4).

Table 4. Pipeline of innovation.

R and D stage in pipeline	Early	Early	Intermediate	Advanced
New, emerging or established	New ideas	New idea advancing an emerging technology but not approved or registered. Potential far from realised	Emerging, and includes new toxins recently approved and registered, with limited field experience. Potential not yet realised	Established potential realised with continuing refinements and improvements Goodnature self-resetting traps
Examples	Trojan females Genetic biocontrol New species-specific toxins	Norbormide Resetting toxin delivery for possums, stoats and rats. Ground and aerial D+ C and PAPP Zinc phosphide for rodents Low dose cholecalciferol	PAPP Zinc phosphide Sodium nitrite Next generation resetting traps New lures and bird repellents	Re-registered 1080 Cholecalciferol Aerial brodifacoum Goodnature traps Deer repellent

A three-pronged approach recognises the need to improve existing tools, complete the development of emerging technologies, and reach out and explore completely new control approaches:

• Prong 1: retaining and improving toxins and traps already in use in New Zealand, and improving how they are used.

• Prong 2: developing new traps and toxins, more effective lures, resetting systems and remote sensing.

• Prong 3: novel approaches for pest control including biocontrol, Trojan females and genetic manipulations.

The focus of the following sections is on recent developments in toxins. There is also a brief commentary on other pest control technologies and strategies, limited to some extent by the small amount of research on resetting traps and novel approaches.

Prong 1: retaining and improving toxins and traps already in use in New Zealand

Toxins

Failure to retain and improve the use of existing toxins and traps will have serious consequences. (Ramey et al. 1994; Eason et al. 1999, 2006, 2011; Adams 2005; APVMA 2008). Cyanide and 1080 have played important roles in mammalian pest control in New Zealand for several decades. Both were reviewed for re-registration in 2006 and 2007 (ERMA 2007), and an intensive independent assessment of 1080 made by the Parliamentary Commissioner for the Environment (PCE 2011). The EPA examined 1080-related research and monitoring covering bait quality and reduced sowing rates, potential sublethal effects, animal welfare issues, secondary poisoning and understanding risk to non-target species, ecotoxicology and fate in water, soil, plants and animals.

Reasons for the reassessment included: 1. that 1080 was registered a long time ago when less toxicology and risk assessment was required; 2. the level of public concern about the use of 1080; and 3. the increased use of 1080 in recent years, mainly for possum control.

Some aspects of the research which underpinned the EPA approval in August 2007 (ERMA 2007) are:

• Quality assurance of baits and reduced sowing rates of 1080.

• Improved understanding of 1080 toxicology.

• Regulatory toxicology to underpin re-registration.

In the early 1990s, high application rates of up to 15 kg/ha of 1080 baits were being used. Considering that only one or two baits were required to kill a possum, these sowing rates were excessive. High sowing rates were primarily used because baits were variable in size, toxicity and palatability, and were distributed unevenly (Morgan et al. 2006). Bait specifications were therefore developed that converted scientific understanding of key parameters (e.g. susceptibility to 1080) into practical manufacturing advice. Once bait of higher quality became available, it was possible to reduce sowing rates and explore more targeted bait delivery patterns (Nugent et al. 2008). Over a decade ago, sowing baits at 3 kg/ha was shown to be as effective as at 10 kg/ha, and even lower rates have been achieved since, although there have been concerns from some practitioners (D. McLean, Project Janszoon, pers. comm. 2016) that this research has gone too far. Ultra-low sowing rates of < 0.5 kg/ha can compromise the success of a 1080 operation, particularly when targeting multiple species, and recalibration is now considered advisable for different habits and circumstances. Similarly, whilst continuing to improve bait quality, better vigilance, audits and good manufacturing practices also need to be sustained.

Important research on the tactical use of 1080, with and without repellents of non-target species (Cowan et al. 2016; Crowell et al. 2016a, 2016b), continues to improve the scale, specificity and intensity of control. Local elimination of possums is being shown to be technically achievable when pre-feeding regimes are optimised, and sustainable when combined with follow-up perimeter control to reduce immigration (D. McLean, Project Janszoon, pers. comm. 2017). Modelling predicts considerable long-term cost savings from this approach (Morgan et al. 2006, 2015; Nugent et al. 2008, 2012). This type of research and new management approaches to best practice with 1080 baiting remains critically important.

The rapid metabolism and excretion of 1080 means that it is unlikely to bioaccumulate in the food chain (Atzert 1971; Eisler 1995), and the ‘withholding period’ for livestock that are suspected of having had contact with 1080 baits can be defined. Prolonged persistence of 1080 in animals after sublethal exposure is also unlikely, and this has been confirmed for animals of all sizes from rabbits to sheep. However, for some reason in the 1990s, rapid elimination of 1080 was confused with safety and lower risk of toxicity. Although 1080 itself is not cumulative, studies in sheep and laboratory rats demonstrate that cumulative damage to the heart or other organs from repeated exposure to sublethal doses of 1080 is possible (Eason & Turck 2002; Foronda et al. 2007a, 2007b), and is more persistent in carcasses (Eason et al. 2012b). Our improved understanding of these risks (Beasley et al. 2009) and potential effects has empowered managers to continue to improve pest control practice to prevent human exposure among those involved in pest control.

Many laboratory-based toxicology studies were completed in the USA before 1995 (Table 2). These included 17 studies on product chemistry, six studies on wildlife hazards and four studies relevant to human health, as summarised in the proceedings of a science workshop on 1080 (Fagerstone et al. 1994; Seawright & Eason 1994). Toxicity data from three different, complementary tests indicate that 1080 is not mutagenic, and further toxicity studies have defined the no observable effect levels (NOEL) for sustained exposure (Eason & Turck 2002; Eason et al. 2011). Improved understanding of 1080 toxicology and regulatory study results reinforce the need for the use of strict safety procedures (Foronda et al. 2007a, 2007b; Beasley et al. 2009). Even so, despite past efforts (Norris et al. 2000), and treatment that can mitigate poisoning, there is still no highly effective antidote (Eason et al. 2011).

Without research of this type, it will be difficult to continue improving the effectiveness of control operations and non-target monitoring needed to meet EPA requirements and community expectations, so as to ensure the future use of 1080 in New Zealand. The EPA allows the continued use of 1080 for the moment, but with additional controls on aerial application summarised as follows:

• a watch list of all aerial 1080 operations to be maintained to enable the EPA to actively monitor all future aerial operations;

• strengthened controls to further mitigate the risks involved in 1080 aerial drops;

• best practice promoted in relation to pre-operation planning, consultation and notification as well as the management of 1080 aerial operations; and

• further research into alternatives for pest control and the environmental effects of 1080.

In 2011 and 2013, the PCE concluded that, to protect our endangered native species, more control of pests was needed using 1080. Research on improvements in management practice and improved community consultation has enabled the continued use of 1080, its re-registration by the EPA in 2007 and endorsement by the PCE. Reducing non-target impacts also requires continued active research and vigilance (Veltman & Westbrooke 2010), and the focus on long-term gains and ecological consequences (Innes & Barker 1999; Innes et al. 2004), are both critically important. Researchers and practitioners of pest control should continue to develop improved baits and delivery, and further advances in strategies and repellents to increase its target specificity as a conservation tool or for TB vector control.

Cyanide was not subjected to an EPA reassessment, but there have been safety concerns over its use in a paste. They have been mitigated by research on the development of an encapsulated pellet form, which eliminates the risk of operator exposure to hydrogen cyanide gas (Eason et al. 2010d; Ross et al. 2011; Shapiro et al. 2011). Neither was cholecalciferol reassessed, because of its low secondary poisoning risk for dogs and cats and low toxicity to birds (Eason et al. 2000, 2010a; Hix et al. 2012).

Anticoagulants such as brodifacoum for rodents and possums, diphacinone for rodents and pindone for rabbits remain essential components of the pest control toolbox, and have been retained with their use refined to improve target specificity. Research continues on brodifacoum, internationally and in New Zealand, to improve its utility, and to address reports globally of extensive wildlife contamination following its field use.

Whilst alternatives to brodifacoum are advancing, optimising the value and potentially greater use of aerial brodifacoum on mainland New Zealand makes sense. Rodent eradication on islands in the past 35 years, using bait in stations and applied from the air, has been spectacularly successful (Taylor & Thomas 1989, 1993; Towns & Broome 2003). The use of brodifacoum to eradicate mammalian pests on larger areas of mainland New Zealand would be a departure from the normal pattern of use, and would require underpinning research and changes to current codes of practice. If it were used behind virtual barrier systems, as advocated by Zero Invasive Predators (ZIP), it could be valuable for one-off eradication rather than continued control, which would mitigate concerns regarding bioaccumulation and social license to aerial broadcast (A. Bramley, ZIP, pers. comm. 2015).

Further research integrating aerial application of 1080 and brodifacoum (which has not received the same attention as 1080) could be fruitful, particularly as aerial application is the most cost-effective tool for small mammal control over large remote areas (Nugent et al. 2012). Future research could see these toxins being more strategically applied by drones with GPS assistance, and these ideas are currently being advanced by Landcare Research, NZ with TBfree and Ministry of Business, Innovation and Employment (MBIE) funding (B. Warburton, Landcare Research, NZ, unpubl. data). In addition, research and management experience in the aerial application of new toxins such as PAPP could be useful, particularly where stoat numbers are high and rodents less numerous.

Traps

Kill traps have been commonly used for decades. As a general rule they have higher public acceptance (http://www.conservation.co.nz/documents/science-and-technical/drds227.pdf) than broadcast poisons (Fraser 2006), and can also be used by community groups and private individuals without the need for special licences or permits. The Animal Welfare Act 1999 and the National Animal Welfare Act Committee have influenced research and trap design, and manufacturers can apply to have their trap assessed for humaneness (Littin et al. 2004).

Recent innovations include improving the design of standard kill traps and live-capture traps, coupled with wireless technology to make monitoring and control along barriers far more effective and substantially cheaper (Jones et al. 2015). These approaches are being pioneered by ZIP Ltd, which has been established as a research and development company to advance new approaches to predator management. A practical ‘remove and defend’ strategy has evolved with intensive use of traps to enable the complete removal of rats, stoats and possums from large areas, and then defend them from reinvasion with a focus on refinements of a ‘barrier’ system and detection of very low numbers of predators when they breach barriers. Research and practical experience provides a platform for extending this approach to larger and larger areas (A. Bramley and D. McLean, ZIP, pers. comm. 2016).

In conclusion, research and management experience continues to deliver considerable gains in terms of the safe and more effective use of established technologies.

Prong 2: new toxins and traps

During the last three decades, the development of alternatives to 1080 and brodifacoum has been driven by: concern that the use of 1080 in New Zealand may be banned (as it has been deregistered or banned in many countries); the persistence of second-generation anticoagulants; animal welfare; and the desire for greater specificity (Eason et al. 2008). Researchers working on new toxins and rodenticides have considered: retrieval of older compounds not previously used in New Zealand; the development of new compounds; the evaluation of natural toxins in New Zealand plants, informed by Vision Mātauranga; and altering first-generation anticoagulants to make them as potent as brodifacoum but without their persistence, bioaccumulation in wildlife or food safety implications.

Zinc phosphide is an example of retrieval of an older compound not previously used in New Zealand but recently approved for use as a possum control agent by the EPA in August 2011 (Eason et al. 2013). Norbormide is a second example which is still the focus of research and has created an interest in genome screening for pest-specific toxin receptor targets (B. Hopkins, Landcare Research, pers. comm. 2015).

PAPP and sodium nitrite are examples of the development of new compounds. Desirable features of these new toxins are: they are lethal to the target species; they are relatively humane; they are orally active and rapidly absorbed; they have relatively short half-lives in blood/organs versus other rodenticides (many of which have long half-lives); they are not persistent in the environment; .they do not lead to secondary poisoning; they have an antidote; they have a reasonable shelf life; and their cost is reasonable. These ‘red blood cell toxins’ exhibit all of these features. They are the first new vertebrate toxins registered for field control of mammalian pests anywhere in the world for > 30 years, and the first designed with humaneness front of mind. However, field experience and research on strategies to make the most of these new toxins and zinc phosphide is still lacking.

Māori have called for alternatives to existing vertebrate pesticides, and one school of thought is that native plants might hold existing natural toxic compounds that could be suitable alternatives to current vertebrate pesticides. After an extensive literature search, and hui with Māori groups and individuals with specialist knowledge, the plant tutu (Coriaria arborea) was selected for further investigation. Toxicity trials have been conducted with tutin, the toxin active in the tutu plant. Tutin was toxic to rats, especially females. It could be a successful rodenticide for Norway rats if a consistent dose of 55 mg/kg could be accomplished. Questions remain over whether this would be technically feasible (Ogilvie et al. in press), and further research would be needed to take it through to registration.

Also worth reconsidering are first-generation anticoagulants with an ultra-low dose of cholecalciferol as an additive. This combination can give first-generation anticoagulants a potency similar to that achieved by second-generation anticoagulants, removing the need to use compounds which are persistent in the environment (Pospischil & Schnorbach 1994). Diphacinone with cholecalciferol (D+C) is the best option as an alternative to brodifacoum where bioaccumulation of residues and non-target effects are a concern (Parmar et al. 1987; Crowell et al. 2013). Bait containing 0.06% cholecalciferol and 0.03% diphacinone has been developed, and dossiers were filed with the EPA late in 2015 to support the registration of a solid bait effective in killing possums and rodents with similar potency to brodifacoum. There is interest in developing an aerial bait containing D+C and, with this in mind, research is being planned by Landcare Research on bait degradation and non-target risks (D. MacMorran, Connovation Ltd, pers.comm. 2017). Time to death in possum is less protracted than for brodifacoum alone, which is important from an animal welfare perspective (Littin et al. 2002, 2004). The key feature of the D+C bait is its similar potency to brodifacoum-containing bait but with a lesser risk of toxin bioaccumulation.

In conclusion to this section, emerging technologies include new toxins with desirable characteristics. Some have been registered and others, such as D+C and norbormide, require further research to enable their registration. Valuable experience, in terms of data requirements and processes has been gained by researchers and regulatory agencies during the registration of PAPP, sodium nitrite and zinc phosphide which should help accelerate approval of new toxins and baits, novel delivery systems and alternative technologies.

To date there has been only limited research and very limited practical experience with their use, especially when compared with over 65 years of refining 1080 baiting. Additional research efforts and equally, if not more importantly, much more extensive practical field experience is badly needed now if the potential of new toxins and other tools to better enable protection of native species from introduced predators is to be realised.

Resetting toxin delivery, more effective lures and remote sensing

New self-resetting delivery systems are under development, building on earlier prototypes (King et al. 2001, 2007a; King 2003). These could allow for up to 200 pest animals to be removed by a single device which also incorporates responsible toxin delivery techniques (i.e. low risk to non-targets). They will complement current control methods and could provide a cost-effective low maintenance solution to reinvasion problems (Blackie et al. 2014, 2016). There are still research and engineering challenges with field reliability. Even then, when fully reliable, they will not always be the most cost-effective option (Warburton et al. 2015), although their species specificity is already a significant advantage.

The Spitfire (Table 5) is designed to attract and control pests over long periods of time with minimal input and maintenance. Preliminary field trials of its resetting toxin delivery system started in earnest in 2013 after several years’ experience with prototypes, and have shown promise. There is a tunnel version for stoat and rat control, and an upright version which attaches to trees for possum control. Both have been field tested (Blackie et al. 2014; Murphy et al. 2014a, 2014b).

Table 5. Author projections and possible future scenarios and timelines for new technology (subject to investment in and beyond 2017). Some of these technologies are advancing with funding; others are temporarily stalled or require modest resourcing. Additional inclusive investment targeting a broader range of researchers with different skills sets, coupled with support for small to medium sized enterprises (SMEs) will be needed to meet these timelines and the aspiration of Predator Free New Zealand 2050.

Impacts	Details
Resetting toxin delivery systems	Final engineering of early prototypes (2018–2021) Extensive or nationwide field trialling the stoat/rat/possums toxin re-setting delivery device (2018–2020) Accumulation of research data on different toxins in these devices, to finalise EPA/MPI registrations in 2021 and optimised field use strategies Incorporation of species recognition technology (2022)
PAPP	Development of effective lower concentration formulation that can be used by community groups without approved licensed operator (2020) Development of long-life PAPP baits (2017–2020) Development of aerial PAPP for stoats and feral cat control over 2017–2023 Accumulation of new research data to support EPA and MPI registration of PAPP baits for aerial control (2018–2024)
ESN and MZP	Extending MZP for rodent control as well as possums in 2020 Developing ESN for delivery in resetting Spitfire systems for humane targeted control of possums in 2021 Field trial experience with farmers and DOC in 2017–2020 of ESN in feral pig control Development of solid ESN and MZP aerial application. MZP could be more effective for both rodent and possum control than 1080 (2018–2022) Stepping-up the accumulation of new research data to support EPA registration of new baits for ground and aerial control
Diphacinone + additive of low dose cholecalciferol (D+C)	Development of D+C in baits that can be effective for both ground and aerial control of rats and possums by 2023 Accumulation of new research data to support EPA and MPI registration of new D+C baits for ground and aerial control (2017–2020) including efficacy, fate in the environment and non-target risks Developing D+C for delivery in resetting Spitfire systems for humane targeted control of possums by 2021 Use in areas where continued use of brodifacoum is a concern
Norbormide	Development of norbormide baits that can be effective for both ground and aerial control of rats by 2020 Accumulation of new research data to support EPA and MPI registration of new norbormide baits for ground and aerial control (2021)
Advanced lures and repellents	Better bird repellents on 1080 baits by 2020 Lures improving resetting devices by 2020
New genetic approaches to mammalian pest control	Complete POC by 2020. Pen and field trials advancing
Barrier and eradication	ZIP extends larger and larger areas using best-of-old and new technologies. Extensive sites larger sites < 1000 ha by 2017–2018

Note: PAPP, para aminopropiophenone; ESN, encapsulated sodium nitrite; MZP, micro encapsulated zinc phosphide.

The tree-mounted possum units are species-specific, lightweight, environmentally robust, and have the ability to kill multiple possums before requiring servicing. They require a possum to stand on a weight-activated platform and simultaneously touch a lured upper trigger, causing a measured dose of a palatable gel containing zinc phosphide to be delivered onto the possum’s abdomen. Pen trial results demonstrated that possums then groomed the gel from their fur, ingested it and died. The first field trial in early 2014 at Cass achieved a 90% kill of radio-collared possums (Blackie et al. 2016). At the second and most recent field trial at Totaranui in the Abel Tasman National Park, a 92% kill was achieved (L. Shapiro, pers. comm. 2017, Boffa-Miskell).

Reducing the potential risks to non-target species was one of the key focuses of the design process, so any species indigenous to New Zealand is very unlikely to trigger the units. This is particularly important for birds such as kiwi and weka (Gallirallus australis), both of which can be at risk with other control methods. At the Totaranui field trial the possum Spitfires remained functioning and reliable for several months in the field, and no weka were killed. It is intended they should be functional for at least 12 months before requiring servicing (Blackie et al. 2016).

Another advantage of the Spitfire for possums is that zinc phosphide does not bioaccumulate in animal tissue and therefore offers low risk of secondary poisoning of non-target vertebrates (e.g. dogs), and it is not persistent in the environment. The delivery system includes the toxin in a sealed cylindrical container, so it also minimises risk to handlers who have no contact with the gel, and there is no spillage or scattering of uneaten bait. After the completion of further field trials in 2015, preliminary registration dossiers for the possum Spitfire were filed with the EPA (Blackie et al. 2016).

The tunnel version of the Spitfire version for stoats uses PAPP, which also will not bioaccumulate and has low risk of secondary poisoning. Based on one field trial to date, registration dossiers for the PAPP stoat Spitfire were filed in 2016 (Murphy et al. 2014a). Spitfires for rats and for multispecies targets are also being advanced in parallel (Murphy et al. 2014b).

Further progress requires overcoming engineering challenges, successful commercialisation of prototypes, completion of registration of toxins and resetting devices for rodents, mustelids, feral cats and possums, and more extensive and vigorous field testing of efficacy in different control and eradication scenarios. Research into optimum spatial deployment strategies, aiming at minimising device spacing and numbers of servicing visits, is still required, and allowance must be made for the absence of any guarantee that an animal finding a control device will engage with it. One trial in 2006 tested the question of minimal spacing by deploying 20 Sentinels (Table 6) loaded with 1080 for ferrets over 2554 ha of farmland (about one per ferret home range). The results confirmed the value of selective (weight-programmable) bait delivery devices, but revealed extensive heterogeneity in the ability of ferrets either to find the widely dispersed devices within the 5 weeks they were available, or willingness to enter the tunnels (11 of 23 ferrets known to be present did not find or enter one) (King et al. 2009). An effective long-distance lure is clearly a critical requirement for any minimal-spacing array, but the cautious behaviour of pest animals towards artificial devices is also a serious issue. Early proof-of-concept field studies in 2004 showed that seven of 13 radio-collared ferrets resident near four observation stations approached within 8 m of a Sentinel and, of the six that approached, only three entered over 8 days (King et al. 2009). These lessons were valuable in informing the cage, pen and field protocols when designing the Spitfire systems, with extensive observations of animal behaviour using video, camera traps and proximity logging systems (Blackie et al. 2014, 2016; Murphy et al. 2014a, 2014b).

Table 6. Approaches that ‘dropped out of the pipeline’.

Name	Target	Mechanism	Rationale for stopping research	Source(s)
STINGER	Possum	Injection of toxin	Inhumane, heavy, costly	Eason & Bacheler (1995)
GOTCHA—toxin tunnel and electronic trap	Stoat	Applied to fur or electrocution.	Tech failings	Department of Conservation (2002)
SET & FORGET—bait station	Possum and rat	Bait station	Lack of appropriate bait	Henderson & Frampton (2004)
SENTINEL—tunnel	Ferret	Selective weight sensor	Failure of commercialisation; converted into platform for the Spitfire	McDonald et al. (1999); King et al. (2007a)
Passive Toxin Application	Ferret	Brushes in tunnel	Limited by not resetting	Fisher et al. (2007)
Virally vectored immune-contraception	Possum	Immuno sterilisation	Lack of efficacy-public attitudes	Tyndale-Biscoe & Hinds (2007). Department of Conservation (2002)
Screening and targeting pest-specific toxin receptor targets	Possum	GABA receptor	Lack of efficacy	Gould et al. (2010)
Red blood cell toxins as rodenticides, including sodium nitrite, PAPP analogues and carbon monoxide releasing molecules (CORMS)	Rat	Red blood cells	Attempting to follow guidance for humaneness (Mason & Littin 2003) but lack of efficacy in rats (versus stoats, cats, pigs and possums)	Eason et al. (2012c); Conole et al. (2014)
Coumatetralyl and cholecalciferol	Possum and rat	Haemorrhage	Palatability and diphacinone having a shorter half -life	This review

New research on lures is underway and is critically important (W. Linklater, Victoria University of Wellington, pers. comm. 2016). Recent work on inter-species communication among mammalian predators has revealed that dominant predator odour triggers ‘eavesdropping behaviour’ by stoats, which has implications for improved lures. Ultra-potent lures to expand the range, efficacy and cost-effectiveness of both resetting toxin-delivery systems and traps could offer ‘transformational change’ in pest control (Campbell et al. 2014), and monitoring of results by site occupancy analysis (King et al. 2007b).

Looking to the future, there are aspirations to combine resetting toxin-delivery systems with species recognition to improve specificity for New Zealand and overseas markets. Analysis of footprint, gait and stride length will allow for mammalian species discrimination by weight and print characteristics as they cross a waterproof, low-cost sensing surface (Irie et al. 2014). Specifications are being extended to detect the prints of animals ranging from much less than 1 gm in weight to 1 kg build on the aspirations of earlier research (King et al. 2007a).

In conclusion research has enabled 1080 to be retained, advanced safer baits and new toxins and delivery systems.

New traps

In the last 10 years, automatic resetting traps have been advanced and improved. Early resetting traps that had engineering, welfare or effectiveness flaws have been discontinued. Goodnature Ltd (Table 4), working with the DOC, has designed new devices to humanely kill animal pests and then reset themselves (Barr et al. 2011; Gillies et al. 2012). Targeted species include stoats, rats and possums. Results obtained in the last 2 years included significant reductions in rat populations at the large-scale DOC trials in Northern Te Urewera, the Boundary Stream Mainland Island in the Abel Tasman National Park. Important recent improvements include a new weka excluder for the rat and stoat trap. Research on lures is also important for increasing the range and cost-effectiveness of resetting traps. Field experience continues to be gained, with some good results (Peters 2015), and the Goodnature A25 for possum control is a considerable advance on earlier versions.

Some recent innovations include improving the design of standard kill traps, coupled with wireless technology to make monitoring and control along barriers simple and effective (Jones et al. 2015). However, many of these examples are still emerging technologies and additional research is needed for them to reach their full potential. One of the most important questions to be resolved is less about control technology than about animal behaviour; namely, the development of behavioural resistance to over-use of artificial devices. In areas subject to continuous long-term trapping, a core population of stoats may learn to actively avoid suspicious-looking tunnels, and thereafter remain untrappable (Veale et al. 2013; Robertson et al. 2016).

In conclusion to this section, significant emerging technologies include new resetting technology which has advanced rapidly in the last few years, but more research is needed for these tools to get them working effectively, coupled with innovation in terms of how they are used to maximise their cost effectiveness and realise their potential. More effective lures could also have a huge impact on their utility.

Prong 3: novel approaches for pest control

Novel control of vertebrate and invertebrate pests, for example the Trojan female technique (TFT), is being explored by researchers at Landcare Research, Monash and Otago University (Gemmell et al. 2013). Natural mutations that cause male infertility have now been identified in the maternally inherited mitochondrial DNA (mtDNA). These mutations have been identified in fruit flies (Drosophila), hares (Lepus europaeus) and mice, and are likely to be widespread in nature. Research aims to harness these mutations through the release of Trojan females carrying the mutations. The sterile male technique is commonly applied to invertebrates, and has eradicated the parasitic screw-worm fly in a number of locations, but requires large quantities of sterile males to be produced and released. Further examples include ‘gene drive’ that enable malaria-resistance to be passed onto mosquitoes’ offspring such that they are unable to transmit malaria (Gantz et al. 2015), and a gene that sterilises all female mosquitoes, which could suppress mosquito populations (Hammond et al. 2015).

Genetic options for the control of invasive fishes, which could be applicable to mammals, were recently reviewed by Thresher et al. (2014). The Trojan Y and several recombinant options that heritably distort pest population sex ratios are deemed technologically feasible, close to proof-of-concept stage and potentially more effective than sterile male release programmes (Gemmell et al. 2013). All genetic options will require prolonged stocking programmes of genetically-modified breeding adults to be effective. Any form of heritable interference with reproductive efficiency would be difficult for mammals, especially stoats (Norbury 2000); but if feasible and socially acceptable, modelling suggests that genetic techniques could be most effective when combined with conventional control options (Thresher et al. 2014): this might also apply for possums, rabbits and rats.

Novel non-lethal and lethal approaches, for example mining of genomes of pest species for species-selective toxin design, sex ratio manipulation and new attempts at fertility control, and reconsideration of mammalian parasites as adjuncts to vertebrate pest control (Tompkins & Veltman 2015) and maintenance of rabbit haemorrhagic disease (Eden et al. 2015), will be important areas for research in the future. Genetic techniques can already be used to quantify reinvasion, survival and in situ breeding rates during control operations (Veale et al. 2013). CRISPR editing capabilities (RSNZ 2016), which were not available to earlier researchers pioneering alternatives to toxins, have considerable potential (Ledford 2016) and should enable research on innovative approaches to advance more quickly.

The pipeline of innovation

A pipeline of technologies has developed through the multidisciplinary efforts of many research groups and management agencies. New ideas and technologies enter the pipeline at an early stage, still requiring fundamental research. Some are emerging as proven technologies close to proof-of-concept, and some are ready to graduate into new tools close to registration and uptake by pest control professional and community groups.

Emergence from the pipeline is illustrated by the example of the three new toxins (Table 4) that have been through the fundamental and applied research steps and registered through the EPA and MPI (ACVM) for field use; namely, PAPP for stoats and feral cats, MZP for possums and ESN for both possums and feral pigs.

There are exciting further developments in the pipeline. These and possible timelines for a selection of new tools are provided in Table 5. To meet these timelines, continued investment will be required beyond 2016 and at present this is uncertain.

There are risks associated with any research, and not all technologies entering a research and development and commercialisation pipeline will succeed (Department of Conservation 2002). Biocontrol in the form of virally vectored immuno-contraception was discontinued after 20 years of imaginative and multidisciplinary research (Tyndale-Biscoe & Hinds 2007). Similarly, earlier attempts to produce a possum-specific toxin based on a unique species-specific receptor were discontinued due to lack of efficacy (Gould et al. 2010). More recently attempts to produce novel rodenticides from carbon monoxide releasing materials and analogues of PAPP as candidate rodenticides (Rennison et al. 2013; Conole et al. 2014) were discontinued as none were sufficiently potent. Early resetting traps or toxin delivery devices (Table 6) have been discontinued.

Conclusions and discussion

New Zealand research has enabled the retention of essential tools such as 1080, cyanide, brodifacoum and traps for broad-scale control and eradication programmes. It has also found ways to optimise their use, and is providing new tools for the control of mammalian pests. Central to the research has been the goal of reversing the decline of New Zealand’s native species through new and improved technologies and tools for pest control and eradication and surveillance. We have also considered replacement of existing tools including 1080. However, regardless of real or perceived deficiencies of 1080 or brodifacoum, these two toxins are essential for both ground and aerial baiting strategies for the foreseeable future, alongside research to refine their use and minimise non-target impacts. There are no ‘silver bullet’ replacements in sight for 1080 (Hansford 2009, 2016).

In 1994 the recommendation of the Royal Society Science Workshop on 1080 was as follows: ‘It is important that the potential adverse effects of existing toxicants and products are continually updated […] so that appropriate code of practice and safety standards are regularly reviewed, alongside new findings to minimise risk […] and alternative safer products are produced’ (Seawright & Eason 1994, p. 164, 168). In terms of existing tools, the last 23 years have seen improvements in understanding the risks and the benefits of the use of 1080, brodifacoum, cyanide and traps in both control and eradication settings. In terms of new tools, the last 6 years have seen a record period for new vertebrate control product registrations.

Two types of toxins for mammalian pest control are still needed in New Zealand, alongside traps for integrated pest management. Comparatively acute-acting compounds include 1080, which remains controversial but still acceptable to the PCE. Brodifacoum is the most widely used and highly effective slow-acting rodenticide, but although its tendency to bioaccumulate and cause secondary poisoning has advantages in some situations, it is not acceptable when repeated control is required or where there are concerns regarding residues and food safety.

A momentum has been generated by Predator Free New Zealand 2050 aspirations to reverse the decline of kiwi and other native species, and to establish predator-free status across large areas of New Zealand. This momentum needs to be accompanied by a continued focus on maintaining and improving pest control tools and strategies, including new rodenticides and other exciting developments with national and global conservation and export potential.

We should learn from what has happened in the past and from overseas where essential tools, capabilities and capacity to make advances have been lost. The continued use of 1080 in New Zealand, without research to further improve its safety and effectiveness, should not be taken for granted. We must also observe carefully any new advances in gene or other technologies being developed in different fields of pest control.

For sustained control options, research is needed to inform the frequency and intensity of control, and the best control tools (Parkes et al. 2017). When comparing control technologies it is important to recognise that some, like 1080 baiting, have been researched and refined at considerable expense for over 65 years. New technologies, whether they be new toxins such as PAPP (Murphy et al. 2007), better traps (Barr et al. 2011), resetting toxin delivery systems or various genetic manipulations (Gemmell et al. 2013) badly need complete research and extensive management experience to be effective resources over the next two decades, and for an appropriate assessment of what is the ‘best option’ or where different technologies might have most benefit.

In this review we have sought to identify toxins and delivery devices that are species-specific, more humane than existing poisons, do not persist in the environment, have a short half-life and do not bioaccumulate (Table 3). Further research and development and field application of these should be complemented by improvements in resetting toxin delivery and resetting trap technology. Similar investment is required, and is already being applied to seed fundamental research on genetic (e.g. TFT and genome screening for pest-specific toxin receptor targets) and other novel approaches to mammalian pest control.

We have the following closing comments:

1. We recommend a three-pronged research strategy aimed at: retaining and improving toxins and traps already in use in New Zealand; developing new traps and toxins, long-life lures, resetting systems, remote sensing, species identification and wireless technologies; and investigating novel non-lethal and lethal approaches for pest control. Anything less puts our national heritage at risk.

2. Research groups involved in the development of new cost-effective technologies to achieve sustainable pest control in pursuit of the government’s 2050 predator-free goals need to include engineers and those individual and commercial partners with track record and ability to advance and register new products or successful biocontrol tools.

3. In recognition of the large number of groups now involved,¹ oversight groups with technical insights are needed to provide research direction and facilitate uptake of emerging technologies. They should include people with a good understanding of technology and research and development processes, coupled with practical field experience and an understanding of current mammalian pest control research nationally and internationally.

4. To date there has been very limited hands-on management experience with the use of new toxins such as PAPP (Dilks et al. 2011), especially when compared with over 60 years of refining 1080 baiting. Additional inclusive investment targeting a broader range of researchers with different skills sets, coupled with support for small and practical experience, is imperative at this stage. Much more extensive field experience, and refinement, is badly needed now if the potentials of new toxins and tools are to be realised.

We add a caution that new technologies will fail to deliver successful restoration if there is no active consultation and collaboration with people who need to use them, especially community-based conservation groups. Biophysical, socioeconomic and cultural constraints will include a range of deeply held and often opposing opinions that need to be identified and acceptable compromises worked out (Norton et al. 2016). Looking ahead (see Table 7), if there is engagement and community support, as well as continuing improvements in the use of existing tools, this multipronged technical approach will facilitate the development and delivery of increasingly humane, species-targeted, low residue pesticides and alternatives to better protect endangered species and reduce disease transmission by mammalian pests.

Table 7. Possible technology scenarios: looking ahead to achieve the predator-free vision.

2020

2030

Better use of 1080 and brodifacoum and traps in a best-of-the-old and new framework An array of improved tools, methodologies and strategies for eradication which are socially acceptable We will better understand pest behaviour at low densities. New detection devices with remote sensing Aerial D+C, PAPP, norbormide and other species-specific toxins emerging Improved repellents of non-targets in 1080 baits. 1080 further refined to deliver 100% kills of possums Resetting traps and toxin delivery devices are being more widely used, but potential still not fully realised. They will be cost-effective and targeted next-generation technologies which have been proved at the small scale to effectively eliminate small mammal pests

Proof-of-concept and advances in non-lethal control After > 10 years outcomes will be socially acceptable, cost-effective and targeted. Next-generation technologies, tools and strategies will be in use over large areas at landscape-scale. Eradication of animal pests will include established lethal control tools and emerging non-lethal technologies. The eradication toolbox will be complete with re-invasion defences in place. Remote wireless detection of re-invading pests. Cost-effective aerial application of target-specific toxins using drones. Fenceless barrier systems over large areas. Technologies established such as species-specific toxins, non-target repellents, pheromones, improved targeting of aerial and ground control tools and linking super-lures with resetting devices and advances in monitoring systems

Acknowledgements

Thanks to MBIE, MPI, DOC, Stefan Endepols (Bayer), Roger Sharples (BASF) and Sharon Hughes (BASF) for supporting research on control technology. Thanks also to the following collaborators and influencers of research and development described in this review: Alan Buckle and Colin Prescott (University of Reading, UK); Peter Savarie, John Eisemann and Kathy Fagerstone (NWRC, USA); Elaine Murphy, Alastair Fairweather, Clare Veltmann, Michelle Crowell and Craig Gillies (DOC); Helen Blackie, James Ross, Ravi Gonneratne, Craig Bunt and Bruce McKenzie (Lincoln University); Steve Lapidge and Simon Humphrys (Invasive Animals CRC, Australia); Duncan MacMorran (Connovation Ltd); Troy Gibson and Neville Gregory (Royal Veterinary College, UK); Joanne Harrison, Gregory Giles and Ivan Sammut (University of Otago); Paul Jansen (consultant, Wellington); Daniel Conole, David Rennison and Margaret Brimble (University of Auckland); Paul Livingstone, Penny Fisher, Tim Marrs, Bruce Warburton, Helen Blackie, Graham Nugent, Dan Tomkins, Richard Bowman, David Choquenot, Campbell Leckie, Don Mackenzie, John Simmons, Kate Littin, Dave Morgan, Les Batcheler, Brian Hopkins, Mark Wickstrom, Lynne Booth, Cheryl O’Connor, Steve Hix, Ray Henderson, Malcolm Thomas, Roger Baldwin, Robert Timm, Colin Rammel, John Reeve, Wayne Temple, Jim Coleman, Phil Cowan, Oliver Sutherland, Devon Mclean, Phil Bell, Al Bramley and many more. Associate Editor: Associate Professor Adrian Paterson.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Regional Councils’ biosecurity managers, TB Free NZ, MPI, MPI ACVM Group, DOC, DOC Pesticide Advisory Group, universities, CRIs, Cawthron, Lincoln Agritech, private companies, Orillion/ACP, ZIP Ltd, Predator Free New Zealand, Conservation Authority, island conservation, pest control professionals, community groups, Biological Heritage NS, MBIE, Ministry of Health, MFE, PCE, EPA.