# Calibration of tunnel tracking rates to estimate relative abundance of ship rats (Rattus rattus) and mice (Mus musculus) in a New Zealand forest > Can footprints in a baited tunnel tell you how many rats are in a forest? In a New Zealand forest, researchers trapped out the ship rats over five nights while counting tunnels with rat footprints, and the share of tracked tunnels fell in step with rat numbers. But as the rats disappeared, mice left more footprints, a reminder that one species can change what another’s tracks tell us. Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/1996-brown-moller-calibration-tunnel-tracking-rates/ ## Citation and access - **Authors:** K. P. Brown, H. Moller, J. Innes and N. Alterio - **Year published:** 1996 - **Journal or source:** New Zealand Journal of Ecology - **Full citation:** Brown, K. P., Moller, H., Innes, J. and Alterio, N. 1996. Calibration of tunnel tracking rates to estimate relative abundance of ship rats (Rattus rattus) and mice (Mus musculus) in a New Zealand forest. New Zealand Journal of Ecology 20(2): 271–275. - **Access status:** Hosted (freely available elsewhere) - **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/1996_Brown_and_Moller_etal_calibration-of-tunnel-tracking-rates-to-estimate-relative-abundance-of-ship-rats-and-mice.pdf ## Tags - **Species:** Mammals > House mouse; Mammals > Ship rat - **Region:** Oceania > New Zealand - **Tracking:** Track counts; Tracking; Tracks; Tracks and signs identification - **Behaviour and ecology:** Behaviour; Habitat selection; Interspecific competition - **Methods and technology:** Chi-square test; Linear regression; Population estimation - **Data type:** New field data - **Observer qualification:** Not reported - **Discipline:** Behaviour and ecology; Wildlife tracking - **Year published:** 1996 ## Detailed keywords - **Species:** Studied: ship rat (Rattus rattus) and house mouse (Mus musculus). Mentioned from other studies: stoats (Mustela erminea, in the tracking tunnel reference), rabbits, wasps. - **Location and habitat:** Kaharoa (37° 57’ S, 176° 25’ E), a forest remnant of about 381 ha, 23 km north of Rotorua, Bay of Plenty, central North Island, New Zealand. Mixed broadleaf and hardwood forest with scattered large podocarps, in a mosaic with scrubland and grassland; vegetation much modified by burning in Polynesian times, browsing mammals and logging. Compared with Rotoehu Forest, a similar central North Island forest. - **Topics:** relative abundance indices; footprint tracking tunnels; tracking rate (percentage of tunnels tracked); calibrating an index against absolute density; removal (extinction) trapping; snap trapping; trappability by age; saturation of indices; effective trapping area and boundary strip; habitat use (forest versus scrub; gully, face and ridge); interference between rats and mice; mouse response to rat removal; ground cover; rodent control monitoring. - **Statistics or models:** Zippin removal method (nightly catch regressed against cumulative catch); simple linear regression (rat density against percentage of tunnels tracked: Density = 0.293 + 0.100 × % tunnels tracked, 98% of variance explained); correlation of mouse tracking rate with rat density (r² = 0.87); chi-square tests; trap success corrected for sprung traps; effective trapping area from a boundary strip based on a 1 ha home range (radius 56 m). Possible wording inconsistencies: the Summary speaks of a linear relationship between rat “trapping and tracking rates”, but the Results test tracking rate against estimated rat density; the Summary gives 6.5 to 7.8 rats per ha as 95% confidence intervals, but the Results say the lower value is the minimum known number (81 rats), and only the upper value is the 95% confidence limit. The Figure 3 caption says “number of rats still alive” while its axis shows density (rats per ha). - **Data type:** New field data: tracking tunnels on 11 January 1994 and on five nights from 21 to 25 January 1994, with removal trapping over the same five nights. ## Synopsis ### 1. What they did In New Zealand, snap traps and footprint tracking tunnels are widely used to give quick indices of how many rodents are about, for example to judge whether rodent control is needed or has worked. These indices assume that a change in the index reflects a real change in numbers, but they had never been checked against true densities of ship rats and mice. The authors set out to do that check, and to see whether removing rats changed how mice used the tunnels. At Kaharoa forest, central North Island, in January 1994 they: - Trapped out the rodents. They laid a 300 m by 300 m grid with 169 rat snap traps 25 m apart, each under a tunnel (natural tunnels of logs and fern fronds, or white plastic or green metal tunnels), baited with peanut butter and rolled oats. They trapped for five nights and used the Zippin removal method, which estimates how many animals remain by plotting each night’s catch against the total caught so far. - Ran tracking tunnels at the same time. Twenty-eight footprint tracking tunnels, baited with peanut butter, were set within a metre of traps, 50 m apart on four lines 100 m apart. They recorded the share of tunnels with rat or mouse footprints on 11 January and on each of the five trapping nights. - Recorded habitat. Each site was classed as forest or scrub, and as gully, face or ridge. - Examined the rats. Rats were dissected for sex, age (adult or juvenile) and breeding condition. They then plotted the nightly rat tracking rate against the estimated number of rats still alive each night, and the mouse tracking rate against rat density. ### 2. What they found Ship rats - 81 rats were caught in an effective trapping area of 12.4 ha. Nightly catches fell quickly: adults 29, 9, 2, 3 and 3; juveniles 12, 9, 11, 1 and 2. - The removal method estimated 83 rats, or 6.7 rats per ha (6.5 to 7.8 per ha). Rats nearly disappeared from both traps and tracking tunnels within five days. - Rat tracking rate rose in a straight line with rat density; the line explained 98% of the variation. There was little if any sign of the index “saturating”, even at a rat density high for mainland forests. - Of 80 rats aged, 56% were adults. Adults were caught more often early on, so they were easier to trap than juveniles. The sex ratio was 45 males to 35 females. Of 19 adult females, 32% were pregnant. - Rats were caught equally in forest and scrub, and equally in gullies, on faces and on ridges. Mice - Mice were caught much more often in scrub than in forest. In forest they were found mostly near tracks with denser ground cover. - Mouse density could not be estimated, because mouse catches went up rather than down during trapping (2, 3, 3, 6 and 6 mice on successive days). - The share of tunnels tracked by mice rose as rats were removed, even though mice were being removed too (r² = 0.87). ### 3. What they concluded - Tunnel tracking rates look like a reliable and cheap way to follow changes in rat numbers in the same place. - The calibration came from one place, one short period and a grid layout of tunnels. It may differ in other habitats, seasons, or with a single line of tunnels, so it cannot yet be used to turn tracking rates into true rat densities. - The rise in mouse tracking after rat removal matches rises reported after rat poisoning. It may be a response to fewer rats (for example, rats excluding mice by predation or scent marking), or mice may have learned that baited tunnels and traps mean food. Further experiments are needed to tell these apart. - If rats do affect how mice use tunnels and traps, comparing mouse indices between places or times with different rat numbers could be misleading, especially where ground cover differs. - The authors urge that removal trapping be paired with routine tracking tunnel monitoring, and that studies like theirs be repeated widely. ### 4. How tracking was or was not used Tracking here meant footprint tracking tunnels: baited tunnels that record the footprints of rodents passing through. The tracking rate was the percentage of tunnels with rat or mouse footprints each night. The paper does not describe how rat and mouse footprints were told apart. No trailing was used. ### 5. Tracker or observer certification Not reported. The paper does not say who read the tracking tunnels or how rat and mouse footprints were identified, although the whole calibration rests on those identifications. ### 6. Other technologies, skills or innovations - Calibrating a footprint index against absolute density obtained by removal trapping. - Three tunnel types over snap traps (natural, plastic and metal) to reduce the risk to non-target species. - An effective trapping area worked out by adding a boundary strip based on ship rat home ranges from radio telemetry in a similar forest. ### 7. Implications for conservation (from the paper) - Rodent tracking and trapping indices are widely used in New Zealand to judge the need for and success of rodent control. This study supports tracking rates as a cheap way to follow rat numbers in one place, but not yet as a way to estimate true density. - Mouse indices may rise after rat removal for reasons other than more mice, which matters when judging the effects of rat control. ### 8. Implications for tracking (Original Wisdom's observations) _These are Original Wisdom's observations, not the authors'._ - The study shows that a simple count of tunnels with footprints can follow real changes in animal numbers closely. It also shows that one species’ presence can change whether another leaves tracks, so the same track count may not mean the same number of animals in every situation. ## Cite this summary Lawrence, K. 2026. Summary of: Calibration of tunnel tracking rates to estimate relative abundance of ship rats (Rattus rattus) and mice (Mus musculus) in a New Zealand forest (Brown and Moller et al., 1996). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/1996-brown-moller-calibration-tunnel-tracking-rates/