Six weanling pigs were given the choice, in a two-sided preference cha
mber, of freely moving between fresh air and ammoniated air (approxima
tely 100 p.p.m. ammonia). The air supplies could be switched between t
he two sides; the two sides were otherwise made as similar as possible
. The pigs were kept in the chamber, one at a time, for 14 days. The f
irst 4 days were a familiarisation period when both sides where suppli
ed with fresh air. During a subsequent 10-day experimental period, the
treatment (fresh air or ammoniated air supply) was changed from side
to side every 24 h. A computer program that monitored the activation o
f pressure pads on either side of a flap door between the two sides re
corded the pigs' movements. During the familiarisation period each pig
established one side on which to spend most of its time; the analysis
therefore tested whether the treatment with ammoniated air changed th
e proportion of the day that each animal spent on that side. The resul
ts show a significant (P less than or equal to 0.001) decrease in the
amount of time spent on the preferred side when that side was supplied
with ammoniated air, with four of the six pigs demonstrating a marked
response. This indicates that pigs may find a contaminant of the air
of piggeries, ammonia, aversive. The potential for detailed study of r
esponses to different substances and concentrations may help to provid
e new specifications for air quality in intensive livestock housing.