The horses for the reserves were carefully selected on criteria of genetic background. All thirteen founders are represented in the Przewalski horses, which were bought from zoos.

A breeding programme could be formulated by means of data from our own data system and computer programmes for founder representation and inbreeding coefficients (BOUMAN 1977 and 1980). Attempts are made to equalise contributions from the wild caught founders and a minimisation of inbreeding.

A total number of eight mares of different genetic background were bought from zoos. The six related stallions, which we had to take with the mares could not be used for breeding. Unrelated stallions had to be obtained separately. New genetic input was achieved by obtaining minimally related stallions from different bloodlines from zoos in the USA, the USSR and Europe for establishing new breeding groups with the female offspring born in the semi-reserves.

In the literature no consensus exists about the level of inbreeding which can be tolerated. Przewalski horses with an inbreeding coefficient of F. 300 and higher produce relatively fewer foals and die younger in captivity (BOUMAN and BOS 1979). The selective pressures in the wild is much more severe. Fitness and a high level of adaptability are therefore required. Only very low inbred Przewalski horses had to be released into the wild. The average inbreeding coefficient of the thirteen foals born in 1989 in the semi-reserves was F. 149, a reduction of 27.3% compared with the average inbreeding coefficient of F. 205 for the 123 Przewalski foals born in zoos and private parks (per 1-1-1988).

Besides the selections on genetic criteria all the Przewalski horses which were offered for sale were inspected first by veterinary surgeons on health, movement and vitality. Three pairs of Przewalski horses had to be rejected after inspection. Two stallions with favourable genetic background could not be used for breeding, because they showed signs of ataxia some months after arrival.

Another guideline of the project was that the Przewalski foals should get the chance to grow up in stable social groups to learn a natural behaviour repertoire from adults and peer groups. Hence we did not rotate adult horses, and we kept the offspring as long as possible in the group, mostly up to two and three years old. The young stallions were released in bachelor groups.

Human contact and interference was kept at a minimum and restricted to health control. In most semi -reserves the Przewalski horses were inspected once a day, in others twice a week. The condition of the horses was observed more carefully every two to three weeks, especially in winter. The climatic conditions in the Netherlands and Germany made it necessary to de-worm them 3-4 times a year. During the first years of their release into semi-reserves the zoo bred Przewalski horses had to be immobilised once a year to cut their hoofs. Their hoofs gradually wore off and broke off themselves, although some horses did better than the others. This could be seen as a very positive adaptation.

Each Przewalski horse was identified by a freeze branded number. Blood typing studies were also performed to aid in identification.

Necropses examinations were performed as a standard. Genetic studies of chromosomes (DIJKSTRA 1984) and studies of the behaviour of the Przewalski horses in semi-reserves were published (FEH 1989, BOEREMA 1984, KEIPER 1989, VAN DER BERG 1990).