Roost type influences torpor use by Australian owlet-nightjars

Lisa Doucette, R Brigham, C Pavey, F Geiser

Research output: Contribution to journalArticlepeer-review

28 Citations (Scopus)


Australian owlet-nightjars (Aegotheles cristatus; ⿼50 g) are one of only a few avian species that roost in cavities year-round and regularly enter torpor. Cavity roosts act as thermal buffers, and roost type likely affects energy expenditure of small birds. We used radiotelemetry to locate diurnal winter roost sites of owlet-nightjars in central Australia and to measure body (Tb) and skin (Tskin) temperature. We also recorded ambient temperature inside (TIN) and outside roosts. Individual owlet-nightjars used one to seven different roosts (tracking time 3-10 weeks), selecting either rock crevices (four birds) or tree hollows (four birds), or switching between the two roost types (seven birds). Rock crevices (TIN +9°C to +33°C) were warmer and thermally more stable than tree hollows (TIN ⿿4.0°C to +37°C). Torpor, often expressed by a reduction of Tskin/Tb by >10°C for 3⿿4 h at dawn, was influenced by roost selection; torpor use in tree hollows was almost twice that in rock crevices. Despite the potential energy savings accrued from roosting in well-insulated cavities, owlet-nightjars roosted in tree hollows more often (65% bird days, n=398) than in rock crevices (35% bird days, n=211). Lower costs of arousal from torpor via passive rewarming and basking and decreased risk of predation are two possible explanations for the preference to roost in tree hollows. We provide the first evidence for the influence of cavity roost selection on torpor use in a free-ranging bird and show that roost selection and thermal biology are strongly interrelated in determining energy expenditure.
Original languageEnglish
Pages (from-to)845-854
Number of pages10
Publication statusPublished - 2011
Externally publishedYes


Dive into the research topics of 'Roost type influences torpor use by Australian owlet-nightjars'. Together they form a unique fingerprint.

Cite this