Like feral horses, the nucleus of the harem of Przewalski horses in the semi-reserves appeared to be the adult mares with their offspring of the past 2-3 years. Social attachment between individual mares and between juveniles were evident at various levels. One Przewalski mare even played a role as godmother to a foal of another mare, defending it when approached too closely by others.

Although ultimately the most dominant member of a harem, mostly the stallion, could influence the group’s activities, initiative and leadership were not shown only by that individual (WARING 1983). It was not always the dominant Przewalski mare or stallion which initiated travelling for instance, but also lower ranking adult mares. The stallion was more often at the rear. The stallion’s role was much more differentiated in the semi-reserves and not restricted to siring mares. The stallion took the leadership role if intruders entered his territory. He always took up a position between his band and the intruder(s).

Herding was seen towards a mare, which could not leave her weak wounded foal. He pressed her to follow the group which had moved away out of sight. The harem stallions were very watchful and were often seen patrolling along the fence in case of disturbances. Although more behaviour data were needed, we got the impression that the Przewalski horses born in the semi-reserves had become very vigilant. Attachment and social facilitation did influence that. Vigilance is one of the basic characteristics associated with zebra predator avoidance together with grouping, male defence of harems and flight (BERGER 1986).

Predator pressure of wolves would be strong for the Przewalski horses after their release into the wild. It was therefore important that Przewalski horses which would be released into the wild had been reared in social groups in semi-natural environments. Suitable  horses for release had to be selected on their life history and behavioural competence.

It was especially important for stallions, on whom group cohesion might depend considerably after release into the wild. The stallions in the bachelor groups of the Goudplaat, had split into two groups. They did bide their time with play-fights and learning to assess the physical capabilities of each other, practising skills they should need later. It is not only strength, age and aggressiveness which play a role, but also the experience of dominance in a bachelor group. They were very watchful in case of disturbances; grouped together in their own band, sometimes both bands grouped together.

All these factors were very important for the selections of suitable animals for re-introduction besides the genetic criteria.