Traps are an essential component of the Remove and Protect model that ZIP is developing to help free the New Zealand mainland from possums, rats and stoats. Under this model, traps are used to (i) defend a predator-free area from incursion by predators and also (ii) detect any individual animals that evade natural barriers (e.g. for possums or rats) or virtual barriers, so that these animals can then be removed before they re-establish a population in the area.

For projects where the goal is to completely and permanently remove predators, it’s essential that the traps used are highly effective. The project goal will never be achieved if too many individual predators can successfully evade the traps, because those predators are likely to breed and establish a population.

Since ZIP’s establishment in February 2015, we have been trying to develop more efficient rat and stoat traps. This Finding describes what we’ve learned over the past four years – including some unexpected events along the way.


Background – the TUN200

The results of a trial run by Hawkes Bay Regional Council at Poutiri Ao ō Tāne between 2011 and 2014 indicated that a single DOC250 kill trap inside in a wooden unbaffled “tunnel” with entry holes in each end caught approximately 50% more ship rats than a standard wooden trap box with a single entrance (i.e. a hole) in one end and a mesh screen at the other end.

Based on the results of this trial, ZIP decided to deploy a tunnel architecture trap for rats and stoats, which we called the TUN200, at our Bottle Rock field site in order to help us research and develop both the virtual barrier and an improved trap.

The TUN200 contained two DOC200s, in order to maximise capture compared with single-set traps, because we were concerned that once a single-set trap was sprung, it might become a significant ‘hole’ in the virtual barrier that we were developing at Bottle Rock. We also thought that a visual lure, positioned between the two traps, was likely to help attract and catch stoats.

The boxes for the TUN200 were made from 12mm thick plywood. In order to 'disguise' the kill trap treadles[1], each of the tunnels leading from the entrance to the centre of the box contained an additional two layers of plywood inside the base to raise the floor to the same level as, or slightly higher than, the treadles.

With the approval of the Department of Conservation we trialled this trap at the Bottle Rock field site[2].

We‘ve used a range of food lures within the TUN200 traps including peanut butter (mixed with icing sugar to thicken it), mayonnaise, Nutella, chocolate, eggs and Erayz[3]. The mouse trap between the DOC200 kill traps shown in the photo was not a standard component of the TUN200 traps.

One issue we identified early on was that the vibration associated with one trap being sprung could cause the second trap to be sprung as well. We were able to significantly reduce such ‘sympathetic spring-offs’ from approximately 30% to <10% by making a 22cm long saw cut in the base of the box, between the two traps, to somewhat isolate them from one another. Spring-offs can also be reduced by placing the trap on a flat surface, preferably soft ground which absorbs some of the impact from the sprung trap.

After extensively using TUN200 traps, we concluded that:

  • it is uncommon for a single TUN200 to catch two rats - rats do not appear to be ‘funnelled’ to traps containing a dead rat (unlike the effect we have observed with possums and leg hold traps)

  • the main advantage of including two traps within a run-through tunnel is that this makes it easy to include a visual lure (such as an egg), compared to within a single-set tunnel – between the two traps where rats cannot access it without first crossing one of the treadles

  • a sprung single-set trap would not be a significant hole in the virtual barrier - particularly when the barrier comprised multiple lines of traps and all traps were being manually serviced (i.e. checked and re-lured) by our field rangers every 3-4 weeks

  • a line of TUN200 traps will catch approximately 40% of the rats that encounter it, as does each successive similar line of traps in a virtual barrier, which is a surprisingly consistent and relatively effective kill rate

Based on these results, we embarked on the process to develop an improved new trap, which we called the ZIP200.


The ZIP200

The fundamental design features of the ZIP200 were that it comprised:

  • an unbaffled tunnel, with similar dimensions to a standard DOC200 trap box

  • a single DOC200 kill trap

  • an automatic safety catch system (for safe servicing of the trap)

  • compatibility with the ZIP-developed automated food lure dispenser and remote reporting system

The ZIP200 trap box was manufactured out of plastic. The design of the trap box also allowed for a visual lure to be placed above the trap plate at an angle that stoats are able to see but not access without first crossing the treadle of the kill trap.

The plastic ZIP200 and a traditional wooden trap box – each for a single DOC200 trap - are shown below.

 In the ZIP Annual Report 2015/16 we reported that other potential improvements to the ZIP200 were likely to include a more open tunnel entrance (to increase the likelihood of interaction, while still excluding non-target species), and a consistent floor and trap plate texture. During the course of the development process, we determined that those features did not improve the catch rate and were therefore not pursued.

The design that we eventually settled upon is shown below.

Once we were happy with the design, we started to test it against the NAWAC (2011) guidelines, at the ZIP Lincoln animal behaviour facility, to ensure that ship rats and stoats caught in the trap would be killed humanely.

During the course of this testing we quickly discovered that some rats were able to escape from the trap after triggering the kill bar. In an attempt to address the fast reflexes of rats and still use the DOC200 kill trap in the ZIP200, we made the kill bar larger. However, some rats still managed to avoid being struck, partly because the larger bar moved more slowly than the standard one. We also tried using stronger springs to increase the speed of the kill bar, but this resulted in the treadle needing to flex more before it was triggered – which in turn exacerbated the sensitivity of rats to that movement.

In total, of 15 ship rats tested, 2 had such speedy reflexes that they were able to avoid being killed by the kill bar that they triggered (like the rat in the video). That’s 13% of the total number of rats tested, which is much too high for us to use the ZIP200 as a trap in projects where the goal is to completely and permanently remove predators.

The Lincoln results prompted us to re-examine the TUN200 trap catch/trigger data from our Bottle Rock field site. The results suggested that for every two TUN200 traps that had caught at least one rat, another TUN200 trap was found where at least one of the DOC200 kill traps had been triggered but was empty (suggesting an escapee rat).

This result led us to wonder whether the designs of the TUN200 and ZIP200 traps were actually worse than the DOC200 in a standard wooden Haines box. To test this, with the support of the Department of Conservation, in late-August to mid-November 2018, we undertook a trial to assess the animal welfare performance for ship rats of the standard DOC200 in a standard wooden Haines box.

Of the 32 rats tested in that trial, over the 60-minute period that each animal was tested for, 37% of the animals interacted with the trap but were not caught by it! That is, they either sprung the trap but escaped without being hit, or entered the box, tested the treadle and then exited the box without springing the trap. These results demonstrate how cautious ship rats can be. That’s not to suggest that these animals would never be caught – our virtual barrier data strongly demonstrates that animals that are not caught on an initial line of traps do get caught on successive lines.

When we assessed the field data and laboratory results we concluded that where the goal is to prevent rats or stoats from establishing a population within a predator-free landscape – rather than suppressing a population to low numbers by catching the majority of rats or stoats – then a more effective rat and stoat trap is required than what currently exists. Consequently, we abandoned our efforts to productionise the ZIP200. However, while this decision marked the end of the road for the ZIP200, we’re already a long way along the path to developing an improved rat and stoat trap.


A new approach

Live capture traps for other mammals are known to be very effective. That’s because, compared to traditional kill traps, live capture traps typically require very little interaction for the animal to be caught and/or the trap ‘disguises’ the element of danger. Consequently, our approach to developing the new trap was to initially focus on designing, testing and refining a trap that is highly effective at simply capturing rats or stoats - i.e. not necessarily on killing them as well. This approach meant that our thinking around the requirements of a highly effective trap to capture predators was not constrained by the design requirements of how to also kill a trapped animal.

Once we decided on the architecture of the live capture components of the trap, we subsequently turned our head towards how to humanely kill a trapped animal. We eventually settled on using carbon dioxide to do this, as this is a standard technique used to humanely euthanise rodents in laboratories.

We also designed the trap to be compatible with our automated lure dispenser and automated reporting system.

We have tested prototypes of the trap at our Lincoln facility, and Bottle Rock field site with encouraging results. We have recently manufactured 500 of the latest prototype, out of plastic, for field trialling. We’ll provide an update about the performance of this trap in the coming months.

Until then, here’s a sneak peek of the latest prototype…

One final note on the DOC200 kill trap

Our decision to develop the latest iteration of the rat/stoat trap was not taken lightly. Designing, testing and refining traps is a time-consuming and costly exercise – which we knew all too well, having invested a lot of effort and money to develop the ZIP200!

We decided to develop this new trap because we’re trying to advance an approach that enables the complete removal of rats and stoats, and prevents them from ever re-establishing populations, at the landscape scale. To do that we need the most effective trap possible. After all, one sexually mature pair of ship rats can theoretically produce 30-120 young per year, and a stoat can produce 8-10 young over the same period.

Where a predator management area is smaller than the landscape scale and/or where the goal is to suppress rat and stoat numbers – i.e. not to completely remove them – then the DOC200 is an excellent trap. Indeed, our testing revealed that the DOC200 in a standard wooden Haines box passed the NAWAC testing for humaneness with a Class A rating. Many threatened native species such as whio and kokako have successfully recovering populations where trapping using DOC200s has been or is being undertaken.


[1] A treadle is the flat steel plate component of a DOC200. The kill bar of a set trap is activated when a rat or a stoat stands on the treadle.

[2] This approval was required because the TUN200 was still being developed, and therefore had not yet been tested against the National Animal Welfare Advisory Committee (NAWAC) (2011) guideline for assessing the welfare performance of kill traps.

[3] Erayz is dried rabbit meat used as an attractant (non-toxic) bait in kill traps.