DOES NEST-HOLE QUALITY, POOR BREEDING SUCCESS OR FOOD DEPLETION DRIVETHE BREEDING DISPERSAL OF TENGMALM OWLS

1. Four hypotheses have been proposed to explain the breeding dispersa l of Tengmalm's owl, a hole-nesting bird of prey which feeds on voles. On the nest-hole quality hypothesis (HQH), the decrease in the breedi ng frequency in old nest-boxes is due to the deteriorating quality of the box. On the predation risk hypothesis (PRH), the decrease in the u se of old boxes is a response to increased predation risk. On the bree ding success hypothesis (BSH), the owls disperse more after poor than good reproductive success. On the food depletion hypothesis (FDH), a b reeding owl pair may reduce vole density in the vicinity of the nest-b ox which increases dispersal. 2. The predictions of the four hypothese s were tested using field data collected in western Finland. In autumn s 1986-87, 28 nest-boxes were replaced by new boxes at the same place on the tree, 28 boxes were transferred and re-erected within 40 - 100 m from the original tree, 27 boxes were both renewed and relocated, an d 79 old boxes were left on the same place on the tree (controls). Bre eding frequency in the boxes was recorded during five years after the treatment. 3. HQH predicts that the breeding frequency should increase when an old box is renewed. PRH predicts that the use of old boxes sh ould increase by relocation. However, the renewal and/or the relocatio n of boxes did not increase subsequent usage. 4. Males were ringed or retrapped at 529 nests (in 1979-91) and females at 655 nests (in 1976- 91). A total of 74 males and 31 females were retrapped as breeders one year later. BSH states that dispersal distances should be negatively related to the breeding success in the preceding year, but this was no t the case. FDH predicts that the owls should disperse less in the inc rease phase than in the decrease phase of the vole cycle. The data wer e consistent with this prediction. 5. The results agreed with the theo retical dispersal models predicting that temporal variation in environ mental quality increases dispersal, whereas spatial variation acts in an opposite way.