# Use of Track Stations to Index Abundance of Sciurids

> Can a length of plastic rain gutter, some chalk and a dab of peanut butter count squirrels as well as weeks of trapping? In Arizona’s pine forests, researchers found that the number of track stations with Abert’s squirrel footprints matched mark-recapture population estimates almost perfectly. The stations were cheaper, lighter and quicker than live traps, killed no animals, and were better at detecting which species were present.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/1998-drennan-beier-use-track-stations-index/

## Citation and access

- **Authors:** Joseph E. Drennan, Paul Beier and Norris L. Dodd
- **Year published:** 1998
- **Journal or source:** Journal of Mammalogy
- **Full citation:** Drennan, J. E., Beier, P. and Dodd, N. L. 1998. Use of track stations to index abundance of sciurids. Journal of Mammalogy 79(1): 352–359.
- **DOI:** 10.2307/1382872
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.2307/1382872

## Tags

- **Species:** Mammals > Abert’s squirrel; Mammals > American red squirrel; Mammals > Cliff chipmunk; Mammals > Golden-mantled ground squirrel; Mammals > Grey-collared chipmunk; Mammals > Rock squirrel
- **Region:** North America > United States
- **Tracking:** Track counts; Tracking; Tracks; Tracks and signs identification
- **Methods and technology:** Analysis of covariance (ANCOVA); Binomial test; Capture-mark-recapture; Detectability; Double sampling; Linear regression; Population estimation; Survey costs
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Wildlife tracking
- **Year published:** 1998

## Detailed keywords

- **Species:** Studied: Abert’s squirrel (Sciurus aberti), golden-mantled ground squirrel (Spermophilus lateralis), rock squirrel (Spermophilus variegatus), grey-neck chipmunk (Tamias cinereicollis) and cliff chipmunk (Tamias dorsalis); chipmunks were analysed together as Tamias. Also recorded in track stations and live traps: red squirrel (Tamiasciurus hudsonicus), on only one plot in 1994. Mentioned: deer mouse (Peromyscus maniculatus), whose light weight meant thick chalk mixtures sometimes failed to record their tracks. One of the two studies indexed squirrels as prey of northern goshawks (Accipiter gentilis).
- **Location and habitat:** Pine (Pinus) forests around Flagstaff, northern Arizona, United States. Eight 60-hectare Abert’s squirrel plots: Pumphouse, Marshall Mesa, Clints Well, Fort Tuthill, Parks, Long Valley, Mormon Lake and Gash Flat. Thirty 2.25-hectare plots in northern goshawk foraging areas.
- **Topics:** track stations (covered track tubes) with chalked aluminium track plates; indices of abundance; validating a track index against mark-recapture population estimates; catch per unit effort (CPUE); presence detection; seasonal change in abundance; statistical power and Type I errors; reference collection of tracks; cost, weight and labour of survey methods; animal welfare (trap mortality); reducing researchers’ exposure to rodent-borne disease such as Hantavirus.
- **Statistics or models:** Mark-recapture population estimates with program CAPTURE (adjusting for edge effects, testing population closure and behavioural responses to capture, choosing the best model); correlation coefficients (r); regression lines compared with analysis of covariance (ANCOVA); multiple pairwise comparisons between plots and seasons, with differences in the two indices tested against a binomial distribution; two-tailed tests. Inconsistencies in the paper: (1) the Methods describe a binomial test for the indices, while the Table 2 footnote calls it a two-tailed t-test of the binomial null hypothesis; (2) the Methods give the 1994 sampling as 24 May to 10 June, while Table 3 says June 1994; (3) Thompson et al. is cited as 1988 in the text but listed as 1989 in the references; (4) the abstract says “trap index” where it appears to mean the track index; (5) the paper is marked as submitted on 9 January 1996, before the April and August 1996 data were collected (probably a revised manuscript, but not explained).
- **Data type:** New field data. Abert’s squirrel plots: live trapping 9 to 20 April and 30 July to 12 August 1996, each followed by two days of track stations. Diurnal squirrel plots: 15 plots from 24 May to 10 June 1994 and 15 plots from 31 July to 9 August 1995, each with one day of track stations followed by live trapping.

## Synopsis

### 1. What they did

Counting animals directly by trapping and marking them is accurate but slow, costly and stressful for the animals. It also exposes researchers to diseases carried by rodents, such as Hantavirus. The authors designed a cheap, rainproof track station for squirrels and tested whether the number of stations that squirrels visited could serve as a reliable index of how many squirrels were there.

- The track station. A covered tube made from plastic rain gutter (25 cm long, 12 cm high, 11 cm wide). Inside sat a removable aluminium track plate, raised on Velcro so rain could run off underneath, sprayed with blue carpenter’s chalk mixed with alcohol, and baited with peanut butter and oatmeal on a small shelf. An animal walking through left chalk footprints.

- A reference collection of tracks. They live-trapped each target species and released them through a runway of six stations. Chalked plates in some stations inked the animals’ feet, and sticky contact paper in the others picked up the prints, which were kept as a permanent reference.

- Testing against population counts (1996). On eight 60-hectare plots near Flagstaff, Arizona, they live-trapped Abert’s squirrels with 144 traps per plot for 6 to 12 days in April and again in August, and estimated each plot’s population by mark-recapture. Right afterwards, they set out 100 track stations per plot for two days and counted how many had Abert’s squirrel tracks. They compared both this track index and the number of squirrels caught in the first one or two days of trapping (catch per unit effort, CPUE) with the population estimates.

- Comparing tracks and trapping for several species (1994 and 1995). On 30 small plots (2.25 hectares) in northern goshawk foraging areas, they set out 100 track stations for one day, then live-trapped on the same grid for one to three days. For each species they counted the stations with tracks and the animals caught.

### 2. What they found

The stations themselves

- One hundred track stations weighed 35 kg, compared with 45 kg for 100 Sherman live traps and 116 kg for 100 Tomahawk traps. Materials cost $1.41 per station (not counting chalk, alcohol and bait).

- Six days of mild rain did not spoil the tracks in any station.

- No animals died in 6,200 track-station days. In live traps, 35 of 1,410 captures (about 2.5%) died, despite midday trap checks.

Abert’s squirrels (1996)

- Mark-recapture estimates ranged from 8 to 45 squirrels per plot in April and 13 to 58 in August.

- The track index followed the population estimates very closely: r = 0.983 in April and r = 0.924 in August. CPUE did nearly as well (r = 0.905 to 0.945).

- The relationship between the track index and population size was the same in April and August (the regression slopes did not differ), so the index held across seasons.

- Both indices picked out differences between plots about as well as mark-recapture, but were considerably less able to show changes between April and August. Mark-recapture showed significant increases on six of the eight plots.

- Each index disagreed with mark-recapture in 5 of 64 comparisons between plots, suggesting a false-difference (Type I error) rate approaching 8%. The authors think the true rate is between 5% and 8%, because mark-recapture estimates have their own biases.

Several squirrel species (1994 and 1995)

- Six species were recorded in both track stations and live traps. Golden-mantled ground squirrels were the most common (59% of stations with tracks, 60% of animals trapped), followed by chipmunks (30% and 35%). Abert’s and rock squirrels each made up under 10%.

- For each species, the number of stations with tracks was strongly correlated with the number of animals caught (r = 0.662 to 0.955).

- Regression slopes did not differ among four species and two time periods.

- Track stations detected a species on seven plots where trapping the next day did not. Trapping detected a species missed by the stations on only two plots.

Effort

In April 1996, estimating one species’ population on one plot by mark-recapture took an average of 15 person-days. Track stations gave indices for four species in one person-day.

### 3. What they concluded

- Both the track index and CPUE are reliable indices of abundance for comparisons within a species, for chipmunks, golden-mantled ground squirrels, rock squirrels and Abert’s squirrels, though the data are conclusive only for Abert’s squirrels. Both are much less statistically powerful than mark-recapture.

- Track counts are a reliable index of seasonal change within a species. The similar slopes across species hint that the index might also compare abundance between species, but the authors do not think their limited study justifies this.

- If budget forces a choice between CPUE and the track index, the track index is favoured for its lower cost and similar power. Lost power could be made up with more plots or more stations per plot.

- Track stations were better than live traps at detecting whether a species was present. The authors suggest animals may be less wary of the stations because they can see through the tube.

- Track stations are cheaper, less labour-intensive, easier to deploy and safer for both the animals and the researchers than live trapping.

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

Tracking was the core of the method. Squirrels left chalk footprints on track plates inside baited tubes, and researchers identified each species from its footprints, using a reference collection of known prints made by live-trapped animals of each target species. The track index was simply the number of stations with a species’ tracks; the researchers did not try to tell whether more than one animal visited a station. Occasionally a station held tracks of more than one species. No trailing was involved.

### 5. Tracker or observer certification

Not reported. The paper does not say who identified the footprints on the track plates or how their accuracy was checked, beyond the use of a reference collection of known tracks. Field workers are thanked in the acknowledgments.

### 6. Other technologies, skills or innovations

- A rainproof track station built from plastic rain gutter, with a removable, raised aluminium track plate held by Velcro and a bait shelf made from halved polyvinyl chloride (PVC) pipe. The design followed suggestions from S. Orloff.

- Chalk and alcohol (one part blue carpenter’s chalk to two parts alcohol) sprayed onto the plate with a garden mist sprayer; it dried within 15 minutes. Thin mixtures coated plates unevenly and thick mixtures sometimes failed to record very light animals, but squirrels were heavy enough to leave good prints with either.

- A permanent reference collection of footprints made by passing live-trapped animals of known species over chalked plates and then sticky contact paper.

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

- Because track stations caused no deaths, they may suit studies of sensitive populations or of the prey base of a sensitive species.

- Track stations reduce researchers’ risk of catching diseases carried by rodents, such as Hantavirus, and avoid the cost and discomfort of the protective clothing and respirators recommended for handling rodents.

- Where budgets are tight, a track index may be an acceptable alternative to mark-recapture, especially if more plots or stations make up for the lower power.

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

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

- This is an early example of testing a track-based index against an independent population estimate rather than assuming it works. Here, simple footprint counts on plates tracked mark-recapture numbers very closely.

- The method rests on correct species identification from footprints, which the paper supported with a reference collection but did not test or report. Where similar species share an area, the identification skill of whoever reads the plates matters to the results.

## Cite this summary

Lawrence, K. 2026. Summary of: Use of Track Stations to Index Abundance of Sciurids (Drennan and Beier et al., 1998). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/1998-drennan-beier-use-track-stations-index/