A preference testing apparatus was used to provide sows with continuous acc
ess to three identical farrowing crates, each with a different floor temper
ature. The concrete floor under each crate contained copper pipe through wh
ich temperature-controlled water was circulated to achieve unoccupied floor
temperatures of 22 degrees C (+/- 3.5), 29 degrees C (+/- 1) and 35 degree
s C (+/- 1). Eighteen sows were tested in the apparatus. Video recording wa
s used to determine sow position from 7 days before farrowing (Days -7 to -
1) to 14 days after (Days 1 to 14). On Days -7 to -1, sows showed no signi
ficant preference among the three temperatures when selecting a resting are
a, Once farrowing had begun, there was a significant increase (P < 0.01) in
the use of the 35 degrees C floor and it became the most preferred resting
area for Days 1 to 3. After this interval, use of the 35 degrees C floor d
eclined significantly (P < 0.01), and use of the cooler floors increased, r
esulting in no significant thermal preference during Days 4 to 6, There was
a further decline in the use of the 35 degrees C floor after Days 4 to 6 (
P < 0.01) to the extent that the coolest floor (22 degrees C) became the mo
st preferred from Days 7 to 14. In summary, sows showed a pronounced increa
se in preference for a warm floor during the 3 days after the start of farr
owing. This change in preference may explain how free-living sows select a
suitable thermal environment for their young, and why sows try to avoid met
al flooring at the time of farrowing. Crown Copyright (C) 2000 Published by
Elsevier Science B.V.