Multiparous sows that had been tethered during lactation were put in two different housing conditions after weaning (Day 0); the sows were either tethered by neck chain, or individually housed in a pen of approximately 6 m2. After two months, ten tethered and eleven loose housed sows were used to assess stress and reproductive parameters. Stereotypic behaviour after the afternoon feeding was assessed from Day 18 onwards; at Day 53 stereotypic behaviour tended to occur during a higher percentage of time in the tethered sows (P = 0.11) and at Day 66, the differences were significant (tethered, 78 +/- 5 vs. loose-housed, 40 +/- 10% of time (mean +/- sem); P = 0.03). At Day 35 and 55, cortisol profiles after afternoon feeding were similar for the two groups of sows (P > 0.10). Around oestrus (approximately Day 64), the profiles of oestradiol-17 beta, luteinizing hormone and progesterone were measured and proved to be similar for both treatment groups (P > 0.10). The duration of oestrus was shorter in the tethered sows (42 +/- 4 vs. 63 +/- 2 h; P < 0.001) and, consequently, the timing of ovulation during oestrus (h after onset of oestrus) was advanced in the tethered sows (28 +/- 2 vs. 41 +/- 2; P < 0.001). The duration of ovulation did not differ (tethered, 2.9 +/- 0.5 vs. loose-housed, 2.1 +/- 0.2 h; P = 0.16). The sows were sacrificed at Day 5 after ovulation; ovulation rate, fertilization rate, embryo development and embryo diversity were similar for the two groups, as were adrenal weight and size of adrenal cortex. Duration of oestrus and the levels of stereotypies at Day 60 tended to be negatively related in the tethered sows (P = 0.07), but not in the loose-housed sows (P = 0.65). In conclusion, sows that had been tethered during pregnancy and lactation, and were housed loose or were tethered again at weaning within two months differed both in stereotypic behaviour and in duration of oestrus, without apparent effects on reproductive hormones.