The spatial dispersion and social interactions were studied in 11 neighbouring pairs of radio-tagged tawny owls Strix aluco in a deciduous wood in Denmark from 1998–2001. The numbers and shapes of territories were stable throughout the survey and similar to a mapping made 40 years earlier. The home ranges of mates were of equal size and overlapped 82% in summer and 56% in winter. The inter-mate distances were on average 2.7% shorter than expected by chance. The activity distribution of neighbouring pairs overlapped 9% (95% CI: 2–15%) on average. Males and females did not differ in overlap with neighbours, and there was a similar overlap between neighbours of the same and opposite sex. Both sexes vocalized more often in the peripheries than in the centres of their territory. The vocal activity during May–September varied extensively among years and months in accordance with variation in the density of juvenile floaters. Males and females vocalized equally often and were involved in disputes with neighbours at similar rates. Usually, neighbouring disputes involved either one individual from each pair or all four. Disputes involving all four owls more often involved chasing and fighting than those involving one owl only from each pair. The dispute rate between neighbouring pairs correlated positively with home-range overlap. The total annual mortality was 21% (95% CI: 6–33%). Dead owners were usually replaced within 1–2 months. Two out of four cases of radio-tagged owls disappearing from their territory because of natural causes was due to take-overs by invading owls, suggesting that the risk of losing fitness resulting from eviction was important. The apparent co-operative territorial behaviour of tawny owl pairs is probably due to improved resource holding potential of pair coalitions compared to single individuals.