The diet of pine martens (Martes martes) with reference to squirrel predation in Loch Lomond and The Trossachs National Park, Scotland

1. What they did

Pine martens are recovering in Scotland, and in Loch Lomond and The Trossachs National Park they have come into contact with introduced grey squirrels for the first time in Britain. Studies in Finland, Sweden and Germany have shown that squirrels can make up a large share of a pine marten’s diet, and there is anecdotal evidence that martens take grey squirrels more readily than red squirrels. The authors set out to find out whether pine martens in the park eat squirrels, and whether that depends on how many squirrels are around.

• Scats. Guided by local knowledge, they collected 30 pine marten scats at each of four sites (Auchineden, Sallochy, Duke’s Pass and Glenbranter) from forestry roads and paths, identifying them with a tracks and signs field guide and recording each location with a handheld Global Positioning System (GPS) unit.
• Diet analysis. In the lab they picked out bones, teeth, hair, feathers, insect parts, seeds and berries under a microscope. Mammal hairs were identified from cuticle impressions and cross-sections, checked against a reference collection that included red and grey squirrel hair.
• Squirrel numbers. At three sites they walked 14 cone feeding transects each (42 in all), counting cones eaten by squirrels to estimate squirrel density.
• Which squirrel. At the same three sites they set 10 hair tubes each for two weeks to see which squirrel species was present.

2. What they found

Diet

• No squirrel remains were found in any scat at any site.
• Mammal remains were field voles, bank voles and wood mice. Field voles were the most frequent prey: in 47% of scats at Auchineden and 37% at Glenbranter, but only 10% at Sallochy and 7% at Duke’s Pass.
• Bank voles turned up in 43% of scats at Glenbranter and 17 to 20% at the other sites. Wood mice were in 10% of scats at Auchineden and Glenbranter and none at the other two.
• Plant material (including rowan berries, seeds and nuts) was in every scat from Sallochy and Duke’s Pass (collected in October and early November), 73% at Auchineden and 17% at Glenbranter.
• Bird feathers and dor beetles were also found.

Squirrels

• Feeding signs suggested no squirrels in the area studied at Sallochy, about 19 at Duke’s Pass (0.08 per hectare) and about 3 at Glenbranter (0.02 per hectare). These are low compared with 0.39 to 0.58 per hectare reported elsewhere in Europe.
• One hair tube at Glenbranter caught red squirrel hair. No grey squirrel hair was found anywhere. Other tubes caught wood mouse hair (Glenbranter) and pine marten hair (Duke’s Pass).

Where the scats were

• Scats containing small mammals were found mostly between 60 and 180 m above sea level.

3. What they concluded

• Pine martens did not appear to be eating squirrels at any of the study sites, in line with earlier Scottish studies. The likely reason is that squirrels were absent or very scarce, and local information suggests squirrel numbers in the west of Scotland had already declined before the study began.
• Field voles, the preferred prey of pine martens in Scotland, seemed plentiful enough that martens had no need to switch to squirrels, which take more energy to hunt. Berries, nuts and seeds may fill in when voles are scarce or when fruit is abundant in autumn.
• Grey squirrels at Sallochy may have retreated to nearby broadleaf woodland, or may have already been reduced by pine marten predation, as local experts have suggested.
• Future diet studies should estimate the numbers of all prey species alongside the diet, and a site where red and grey squirrels live together with pine martens could show whether martens prefer one squirrel species.

4. How tracking was or was not used

Tracks and signs were central to the study. Pine marten scats were found along forestry roads and paths and identified with a tracks and signs field guide (Bang and Dahlström 2001). Squirrel numbers were estimated from feeding signs: cones stripped by squirrels along transects. Hair left in hair tubes, and hair inside scats, was identified under the microscope. No trailing was reported.

5. Tracker or observer certification

Not reported. The paper does not say who identified the scats or how their skill was checked; the fieldwork was a student dissertation project. This matters because no DNA check was used, so the diet results rest on scats being correctly identified as pine marten. The hair identification method, by contrast, was checked against a reference collection of known hairs.

6. Other technologies, skills or innovations

• Handheld GPS (eTrex) to record scat locations.
• Microscopic hair identification: gelatine cuticle impressions viewed at 400x, and thin cross-sections cut from hairs set in glue on balsa wood, compared with museum and trapped reference specimens.
• Cone feeding transects and hair tubes, following Forestry Commission squirrel survey methods (Gurnell et al. 2001).

7. Implications for conservation (from the paper)

• Helping pine martens recover may conflict with red squirrel conservation. Martens eating grey squirrels would help red squirrels by removing a competitor and the risk of squirrel pox, but an earlier study found that where martens had been present for over ten years, red squirrel numbers and recruitment were lower.
• Keeping small mammal populations healthy, for example by including rough grassland or more open forest areas in land management plans, could give martens plenty of voles and possibly reduce the risk of them turning to red squirrels.