# Utilizing a multi-technique, multi-taxa approach to monitoring wildlife passageways in southern Vermont

> Do wildlife underpasses really work? On a new highway in Vermont, researchers combined marble-dust track beds, cameras, snow-tracking, live traps and road kill counts to see not only which animals used the passageways, but which crossed the road instead or turned away. Deer and mink took to the structures, coyotes mostly crossed over the road, and frogs and toads made up a large share of the road kill.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2013-bellis-griffin-utilizing-multi-technique-multi/

## Citation and access

- **Authors:** Mark A. Bellis, Curtice R. Griffin, Paige Warren and Scott D. Jackson
- **Year published:** 2013
- **Journal or source:** Oecologia Australis
- **Full citation:** Bellis, M. A., Griffin, C. R., Warren, P. and Jackson, S. D. 2013. Utilizing a multi-technique, multi-taxa approach to monitoring wildlife passageways in southern Vermont. Oecologia Australis 17(1): 111-128.
- **DOI:** 10.4257/oeco.2013.1701.10
- **Access status:** Hosted (freely available elsewhere)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/2013_Bellis_and_Griffin_etal_utilizing-a-multi-technique-multi-taxa-approach-to-monitoring-wildlife-passageways.pdf

## Tags

- **Species:** Mammals > American mink; Mammals > American red squirrel; Mammals > Bobcat; Mammals > Coyote; Mammals > Deer mouse; Mammals > Domestic cat; Mammals > Eastern chipmunk; Mammals > Fisher; Mammals > Grey fox; Mammals > Long-tailed weasel; Mammals > Meadow vole; Mammals > Muskrat; Mammals > North American river otter; Mammals > Northern short-tailed shrew; Mammals > Raccoon; Mammals > Southern red-backed vole; Mammals > Stoat; Mammals > White-footed mouse; Mammals > White-tailed deer; Birds > Wild turkey; Mammals > Woodchuck
- **Region:** North America > United States
- **Tracking:** Gaits; Scat; Sign surveys; Snow tracks; Track counts; Track measurements; Track photographs; Tracking; Tracks; Tracks and signs identification
- **Behaviour and ecology:** Behaviour; Connectivity; Movement; Road crossings; Wildlife corridors
- **Methods and technology:** Camera traps; Capture-mark-recapture; CyberTracker; CyberTracker software; Descriptive statistics; Global Positioning System (GPS); Pearson correlation
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 2013

## Detailed keywords

- **Species:** Studied with track beds and cameras in the passageways (target species named in the methods): white-tailed deer (Odocoileus virginianus), moose (Alces alces), black bear (Ursus americanus), bobcat (Lynx rufus), red fox (Vulpes vulpes), coyote (Canis latrans), North American river otter (Lutra canadensis), raccoon (Procyon lotor), Virginia opossum (Didelphis virginiana), striped skunk (Mephitis mephitis), long-tailed weasel (Mustela frenata), ermine (Mustela erminea), fisher (Martes pennanti), American mink (Neovison vison) and woodchuck (Marmota monax). Recorded crossing the structures on track beds and cameras: white-tailed deer, woodchuck, raccoon, bobcat, wild turkey (Meleagris gallopavo) and domestic cat (Felis silvestris catus); also muskrat (Ondatra zibethicus) and mink in 2005; frogs (Ranidae), jumping mice (Zapodidae) and Peromyscus mice on the track beds. Snow-tracked: coyote, bobcat, mink, fisher, long-tailed weasel, river otter, gray fox (Urocyon cinereoargenteus), raccoon, white-tailed deer and domestic cat. Live-trapped: deer mouse (Peromyscus maniculatus) and white-footed mouse (Peromyscus leucopus), recorded together as Peromyscus spp., southern red-backed vole (Clethrionomys gapperi), eastern chipmunk (Tamias striatus), jumping mice (Zapodidae), meadow vole (Microtus pennsylvanicus), northern short-tailed shrew (Blarina brevicauda), red squirrel (Tamiasciurus hudsonicus), long-tailed weasel and ermine. Road kill: frogs and toads (Anura), snakes, turtles, white-tailed deer and other animals grouped by size; birds were left out. In the area but not detected in the passageways in 2005: bobcat, coyote, fisher, otter, North American porcupine (Erithizon dorsatum) and American beaver (Castor canadensis). Mentioned as examples elsewhere: elk (Cervus canadensis), Florida panther (Puma concolor).
- **Location and habitat:** The Bennington Bypass (Highway 279), a 7 km two-lane highway from Hoosick Falls, New York, to Bennington, Vermont, United States, opened in October 2004. Two “extended bridge” wildlife passageways over East Airport Brook (43.3 m long, 18 m high; an intermittent stream) and West Airport Brook (56.55 m long, 12.17 m high; a perennial stream), 0.6 km apart, both flowing into the Walloomsac River, with 2.5 m chain-link lead fencing at each entrance; plus a 1.65 m wide, 124 m long drainage culvert linking two retention ponds. Northern hardwoods forest of American beech (Fagus grandifolia, spelled “grandfolia” in the paper), maple (Acer spp.) and eastern hemlock (Tsuga canadensis) with a Canada honeysuckle (Lonicera canadensis) understorey; a 15 m mowed right-of-way on both sides; rolling hills about 4 km west of the Green Mountains; sparse residential housing; mostly private land. Average daily traffic 4,290 to 7,578 vehicles.
- **Topics:** road ecology; wildlife crossing structures and underpasses; monitoring the effectiveness of highway mitigation; a conceptual model of nine possible animal movements near a road and passageway (crossing the road, collision, approaching and avoiding fencing, structure or road, passing through the structure); road as a barrier to small mammal movement; lead fencing and highway fencing; road crossings by coyotes and deer; landscape structure and deer crossing points; amphibian road mortality; setting goals and success criteria for mitigation; comparing monitoring methods (advantages and disadvantages of track beds and camera traps); snow-tracking as a low-cost alternative to telemetry.
- **Statistics or models:** Mostly descriptive counts, proportions and indices: track sets as an index of structure use; road kill index (road kills per survey) by size group, month and distance from the structures; straight-line distances between small mammal recaptures. Pearson’s correlation between distance from the structures and road kill (r = -0.644, p = 0.167). P values are given in Tables 3 and 4 without naming the test used.
- **Data type:** New field data. Track beds: 2005 to 2007 (2005 a trial season; data from 2006 and 2007 used). Cameras: 143 days in 2006 and 133 days in 2007; stream cameras 24 May 2006 to 8 October 2007. Snow-tracking: 24 surveys (47 track nights), 11 December 2005 to 25 February 2007. Small mammal trapping: 48 sessions, 31 May to 17 October 2006 and 8 June to 17 October 2007. Road kill surveys: 21 June to 26 August 2005, 14 April to 16 October 2006, 24 April to 15 October 2007. Paper inconsistencies: (1) the snow-tracking Results say 57 passage crossings and 68 road crossings, but Table 2 and the Discussion give 68 structure crossings and 57 road crossings; (2) track bed monitoring is given as “128 track nights in 2007 and 84 track nights in 2008”, while Table 1 gives these for 2006 and 2007, and the Results report 786 track sets “over 349 track nights for the three field seasons”; (3) the Results say 69% of small road-killed animals were unidentifiable and 31% were frogs and toads, while the Discussion says “only 69% of our small animals were identifiable” and calls 31% a share of all road kills; (4) the abstract says one small mammal crossed the road, the Results call it “one possible road crossing”; (5) the Table 2 caption gives January 2006 to February 2007, the text 11 December 2005 to 25 February 2007; (6) coyote road crossings are given as 31 (movements a and c) in the Results and 29 (movement a only) in the Discussion, where a sentence also says “a still higher proportion of coyotes were found moving across the highway (movements e and f)”, which are passageway movements; (7) the abstract on the publisher’s web page differs from the one in the PDF (for example “through 2005, 460 terrestrial and 300 aquatic crossing structures” versus “through 2008, nearly 700 terrestrial and 10,700 aquatic”).

## Synopsis

### 1. What they did

Roads take away and break up habitat, kill animals and get in the way of their movement. Many wildlife crossing structures have been built in North America, but few are monitored, and most monitoring only records which animals use the structure, not those that cross the road instead or turn away. The authors set out to monitor a whole range of animal movements, for many kinds of animals, around two wildlife passageways on a newly built highway, the Bennington Bypass in southern Vermont.

They drew a conceptual model of nine possible movements an animal can make near a road and passageway, from crossing the road safely or being hit, to being guided by fencing through the structure, to approaching and turning away. They then used five methods together and sorted their records by these movements:

- Track beds and track plates. Boards covered with a fine layer of marble dust were laid across both passageways, and sooted aluminium plates were set in the drainage culvert. Track sets were identified to species (or at least family) with their direction of travel. Each track set counted as a track bed crossing, used as an index of how much the structure was used.

- Camera traps. Film and infrared digital cameras at the track beds confirmed species, and digital cameras along the streams recorded animals passing where there were no track beds.

- Snow-tracking. After snowfalls, observers walked a grid of transects parallel and perpendicular to the highway, covering 500 m beyond each passageway, and recorded every set of tracks and what the animal did in relation to the road.

- Small mammal trapping. Live traps on lines beside both structures, with ear tags, to see whether small mammals moved through the structures or over the road.

- Road kill surveys. Three drives a week along the whole 7 km, recording every carcass by size group and location.

### 2. What they found

Track beds and cameras

- The track beds recorded 786 track sets from at least 26 kinds of animals. Of these, 110 could only be called small or medium-sized mammals, and 62 could be identified only to family or genus.

- The stream cameras recorded 90 crossings by six species: white-tailed deer (53), wild turkey (12), bobcat (9), raccoon (9), woodchuck (6) and domestic cat (1).

- Adding the camera records raised the number of structure crossings detected for these six species by 38% in 2006 and 41% in 2007. Bobcats were detected only by cameras, and deer sometimes walked around the track beds (for example at East Airport Brook in 2006, with no deer on the track bed but 21 on camera).

Snow-tracking

- 162 track sets over 24 surveys, from 10 species: coyote, bobcat, mink, fisher, long-tailed weasel, river otter, gray fox, raccoon, white-tailed deer and domestic cat.

- Nine of the 10 species used the structures and seven crossed over the road. Mink and otter, which travel along streams, only used the structures.

- Overall, 68 crossings were through the structures and 57 over the road (Table 2 and Discussion; the Results text gives these the other way round).

- White-tailed deer made the most structure crossings, and these rose by 75% from the first winter (12) to the second (21), even with fewer track nights.

- Coyotes mostly crossed over the road rather than through the structures, but 15 of their 54 identified movements showed them turning away from the road or the structures.

- Only 17% of animals that came to the structures (25 of 143) turned away from them.

Small mammals

- 690 small mammals were ear-tagged over 48 trapping sessions; 92% were Peromyscus mice. Just over half (55%) were recaptured at least once.

- 15 individual Peromyscus mice made 26 crossings through the structures, and one possibly crossed the road. No other species was recorded crossing.

- More than 36% of recaptured Peromyscus mice moved 65 m or more, the shortest distance from one forest edge to the other through a crossing structure.

Road kill

- 1,289 road-killed animals were counted in 148 surveys. Three quarters were small animals, and at least 31% of the small animals were frogs and toads.

- Very few deer were killed (about 0.05 to 0.07 per survey), though many deer lived in the area.

- Traffic and the road kill rate both rose each year, and road kill did not drop closer to the crossing structures (no significant correlation with distance).

### 3. What they concluded

- The Bennington Bypass structures are used often and by many species, but animals still cross over the road and risk being hit. Bobcats and coyotes seem to cross where pre-existing game trails meet the road at least as often as they use the structures.

- The large structures suit medium and large animals, especially deer, but their lack of vegetation cover may hold back small mammals. This may change as the area grows back.

- Whether the structures “work” depends on the goal: if it is to keep animals off the road, they are not fully effective; if it is to let part of each population cross, they seem effective for some species, such as deer and mink.

- The lack of high fencing along the highway may explain the many road crossings.

- Track beds and cameras used together give a good index of structure use but cannot tell individuals apart. Mark and recapture or telemetry is needed for that. Snow-tracking shows movement in the wider landscape but only in winter.

- Monitoring a broad range of movements and taxa, ideally combined with population modelling, gives a more accurate picture of whether mitigation succeeds.

### 4. How tracking was or was not used

Tracking was central to two of the five methods.

- Track beds and plates: fresh marble dust on boards inside the passageways, and sooted plates in the culvert, recorded the footprints of animals passing through. For each track set the team recorded species (or family) and direction of travel. Difficult tracks were photographed and measured (foot width and length, stride and straddle) for later identification. Track beds could not be read after rain or wind, and small mice running up and down the beds made true crossings hard to judge.

- Snow-tracking: 48 hours after a snowfall of more than 1.3 cm, observers walked the transect grid and recorded species, track and gait measurements, gait pattern, direction of movement, scat and scent marking, where animals crossed the highway, weather, days since the last snowfall and snow depth. Each track set was then classed by the movement it showed in relation to the road. Deer and domestic cat tracks were so many that individual movements could not be told apart, so only their crossings were counted.

The authors list the strengths and weaknesses of track beds (cheap, working round the clock, simple and easy to spread over wide areas; but high upkeep, weather-sensitive, often hard to tell species apart, unable to tell individuals apart, and needing a flat surface) and describe snow-tracking as a low-cost alternative to telemetry.

### 5. Tracker or observer certification

Not reported. The paper does not say who identified the tracks or what tracking experience they had, although the track bed and snow-tracking results rest on track identification and on reading the direction and type of each movement. Four field technicians are thanked in the acknowledgements.

### 6. Other technologies, skills or innovations

- Palm Pilot handhelds running CyberTracker software linked to Global Positioning System (GPS) receivers, used to record snow-tracking data in the field.

- TrailMaster TM1050 active infrared 35 mm film camera, and Reconyx Silent Image PM35M13 motion-sensing infrared digital cameras, set to take 10 images per trigger.

- Track beds of oriented strand board (OSB) sheets covered with about 2 mm of marble dust; track plates of aluminium sheet sooted with a propane torch, with white contact paper.

- Sherman live traps (226) and metal ear tags for small mammals.

- A conceptual model of nine possible movements near a road and passageway, used to compare results across methods.

### 7. Implications for conservation (from the paper)

- Monitoring only what uses a crossing structure misses animals that cross the road or turn away, so it can overstate how well mitigation works.

- Clear goals and success criteria are needed when planning mitigation: keeping large animals such as moose or elk off the road calls for different measures than letting enough animals cross to keep populations connected.

- Highway fencing is likely needed if the goal is to prevent collisions and steer animals into the structures.

- Amphibians made up a large share of road kill, so monitoring should cover many taxa, not only large mammals.

- More information on population size and trends is needed to judge how many crossings each species needs and how road kill affects its survival.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The study shows how tracks and signs can answer questions that cameras alone cannot: snow-tracking revealed not just which animals passed through the structures, but which crossed over the road, followed the fencing, or turned away.

- Cameras and track beds each missed animals the other caught: deer stepped around the track beds and were seen on camera, while the track beds picked up many small and medium animals the cameras missed. Combining tracks and signs with cameras gave a fuller picture than either alone.

- The difficulty of telling individual deer and cats apart in heavy snow-track traffic, and of reading true crossings from mice running back and forth on the beds, shows where the limits of track interpretation lie in this kind of monitoring. Because the paper does not report who read the tracks, readers cannot judge how reliable the track identifications were.

## Cite this summary

Lawrence, K. 2026. Summary of: Utilizing a multi-technique, multi-taxa approach to monitoring wildlife passageways in southern Vermont (Bellis and Griffin et al., 2013). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2013-bellis-griffin-utilizing-multi-technique-multi/