Seminatural environment

The seminatural environment (SNE; 240 × 210 × 75 cm) consisted of two parts: an open area and a burrow system (Fig. ​(Fig.2;2; Chu and Agmo 2014; Houwing et al. 2019a; Snoeren et al. 2015). Four openings (8 × 8 cm) connected the two areas. In the open area, two partitions (40 × 75 cm) simulated natural obstacles. The burrow system consisted of connected tunnels (width 7.6 cm, height 8 cm) and four nest boxes (20 × 20 × 20 cm). Plexiglas covered the burrow at the height of 75 cm, while the open area remained open. A curtain between the two parts allowed for different light settings. The burrow was left dark the entire time. In the open area, on the other hand, light settings simulated a day-night cycle. A lamp located 2.5 m above the floor, simulated daylight (180 lux) between 22.45 and 10.30. From 10.30 to 11.00, the lights gradually decreased to 1 lux (simulating moonlight). The darkness lasted until the light gradually increased from 1 to 180 lux between 22.15 and 22.45.

The Seminatural Environment. Illustration of the whole seminatural environment (A) and sectioning of the different locations (B). 1 = open area close to burrow left, 2 = open area close to burrow right, 3 = open area far away from burrow left, 4 = open area far away from burrow right, 5 = tunnels far away from open a, 6 = tunnels close to OA, 7 = nestbox left, 8 = nestbox mid-left, 9 = nestbox mid-right, 10 = nestbox right. Created with BioRender (https://biorender.com/). A picture of the seminatural environment can be found in the Supplemental materials (Figure S1).

The whole ground of the SNE was covered with a layer (2 cm) of aspen wood chip bedding (Tapvei, Harjumaa, Estonia). The nest boxes had 6 squares of nesting material in each (non-woven hemp fibers, 5 × 5 fibers, 5 mm thickness, Datesend, Manchester, UK). Three plastic shelters (15 × 16.5 × 8.5 cm, Datesend, Manchester, UK) were placed in the open area. Additionally, 12 aspen wooden sticks (2 × 2 × 10 cm, Tapvei, Harjumaa, Estonia) were randomly placed around in the SNE. A pile of food pellets (approx. 2 kg) and four bottles of water were available at all time (see location in Fig. ​Fig.2A2A).

Two video cameras (Basler) were mounted on the ceiling, 2 m above the open area (regular camera) and the burrow system (infrared camera) respectively. Media Recorder 2.5 was employed for video recordings. The data got immediately stored on an external hard drive. The recording was manually stopped and restarted every 24 h. The purpose was to ensure that eventual errors only would affect one day of recorded data.