When to mix sows in groups?

In North America stall housing predominates, however some individuals and larger production companies have been incorporating loose housing in both new builds and renovations. In response to this our research programme looks at both traditional stall and loose housing management systems. Scientific research has demonstrated that sows housed in groups can perform as well as those in stalls, this combined with increasing consumer and retailer pressure to limit use of confinement systems has led to the banning of gestation stalls in many parts of the world.

In those countries where group housing is required, it is generally stipulated that sows must be placed into groups by five weeks gestation. Housing sows in breeding stalls until confirmation of pregnancy at around 28 days post breeding is permitted, as this facilitates individual feeding and avoids mixing aggression during critical periods of the oestrous cycle and embryo implantation. However, pressure continues for the total elimination of close confinement: keeping sows loose at all times during gestation, and potentially even during farrowing. Already, a number of EU members require reduced stall use, with the Netherlands requiring sows to be out of stalls after four days post-insemination. Grouping sows post-insemination is known to work well, provided that mixing occurs rapidly thereafter and the period of social stress does not overlap with embryo implantation.

You may also find interesting:


Group sow housing; coming to America


For the US swine industry, the most controversial animal welfare issue debated over to date is the discussion on housing facilities of pregnant sows; more specifically, the use of gestation stalls/crates.

Producing totally stall-free pork requires a different approach to sow management around breeding, and has some potential benefits as well as some risks. Mixing sows at weaning will give animals time to establish their social group before they cycle, and thus avoid any effects of mixing stress on sow conception rate. In terms of the effect of early mixing on oestrus, it has been suggested that early mixing may disrupt the onset of oestrus in sows.
However, there is research to suggest that if the mixing stress is acute it can stimulate a quicker return to oestrus in sows. Through allowing sows to display oestrus behaviour, there is also the potential to have a greater synchronisation of oestrus within the breeding group. These management options need to be investigated to determine what is best for the sow, her welfare and productivity, and in turn whether there are any other advantages to be gained.

Is grouping sows at weaning viable?

The study, performed in collaboration with Dr Yuzhi Li from the University of Minnesota, United States, compared the effects of mixing sows at weaning, with mixing at five weeks gestation. Measures were collected to evaluate effects of the treatments on sow aggression, welfare and productivity. Three treatments were evaluated:

• Early Mixing (EM): Sows mixed into groups directly at weaning;
• Late Mixing (LM): Sows stall housed at weaning, and mixed at five weeks gestation;
• Pre-socialisation (PS): Sows mixed for two days after weaning, then stall housed for breeding and up to five weeks gestation, after which sows were remixed into the same groups.

The PS provided an intermediate treatment to examine the interaction between mixing at critical time points in combination with housing sows in stalls during the implantation period and further determine if there would be any benefits, such as reduced aggression in the second mixing, if the social group has already been formed at weaning. For all treatments, sows (14 per pen, 2.2 m²/sow) were housed in fully slatted group pens, with partially slatted free access stalls. The stalls were used to house sows during feeding, heat checks and breeding. When the treatment required that sows be loose in a group, the sows were fed each morning in the free access stalls, after which they were locked out of the stalls for up to 22 hours per day in the communal loafing area. When treatments required that sows be kept in stalls, the sows were locked into the free-access stalls.

Sow aggression, welfare and reproductive performance (wean-to-service interval, conception rate, and farrowing performance) were measured. Additionally, oestrus behaviour was measured in the EM groups to determine levels of the behaviour and effects on sow injury or production.

Results: Each system can work

Aggressive interactions on the two days following mixing were no different across treatments, and overall levels of aggression were low. Similarly, no differences were found in sow cortisol levels or lameness. Skin injury scores were lower in PS sows compared to EM and LM sows after the first mixing. However, injury scores on all sows were very low. These data indicate that sow welfare was not significantly affected by the different mixing treatments.

Oestrus behaviour, including mounting, attempted mounts and flank nosing, was observed in EM groups on days 3 and 4 after weaning. The average percentage of sows per pen involved in oestrus behaviour rose from 30% on day 3 to 48% on 4.

Sows in the EM treatment had the highest conception rate (98%), and LM the lowest (87%), with the PS treatment in between (Table 1). It is not clear why the LM treatment, which is the standard practice for managing sows and typical in this herd, had the lowest conception in this trial. It may reflect sub-optimal stimulation of oestrus in stalls when compared to the EM and PS groups, which received mixing stress immediately at weaning. There is evidence that correct timing of stress post-weaning can bring on oestrus, and thus these treatments may have stimulated follicular growth and enhanced oestrus expression.

Overall measures of production were not statistically significant between treatments. However, the EM treatment had a significant reduction in the number of stillborn piglets. The reduction in stillborns appears to indicate a beneficial effect of allowing sows free movement during the early stages of pregnancy. This may be an effect of improved sow fitness, or could be related to embryo placement along the uterine horns, and subsequent placental attachment. Related studies in human health have shown that maternal physical activity can influence placental development and viability.

What can be concluded?

Under good conditions of management, where sows are housed in static groups and individually fed, grouping sows at weaning does not negatively impact sow performance or welfare. Moreover, there may be production advantages to mixing sows into groups at weaning, as indicated by the improved conception rates and reduced stillborns found in this trial. These effects should be explored further. The same results may not be true in a group feeding system in which sows have to cope with greater levels of competition, such as in a heavily stocked ESF system, or competitive feeding system (e.g. floor feeding). While no effect on sow lameness was found in this trial, pens in which sows are mixed should have good quality flooring to reduce injury, and particularly if expression of oestrus behaviour is anticipated. In conclusion, grouping sows at weaning is a viable option under the correct conditions of management.

This research was funded by the National Pork Board, supported by funding to the Prairie Swine Centre from the Saskatchewan Agricultural Development Fund, Sask Pork, Manitoba Pork, Alberta Pork and Ontario Pork.

By Dr Yolande Seddon & Dr Jennifer Brown, researchers in swine ethology and welfare, Prairie Swine Centre, 
Saskatoon, SK, Canada