In outbred population of white rats (Rattus norvegicus), we picked out two groups of male animals with high (HA) and low levels of anxiety (LA). Social preference for a familiar or unfamiliar conspecific were studied. The rats (n = 85) were housed five per cage for two months before the experiment. Thus, the social structures in every cage were stable. The anxiety was tested in an elevated plus-maze. For social interaction test we used the box, in which two opposite corners were separated with perforated transparent Plexiglas walls. A mate of a tested subject was placed into one corner the cage, an unfamiliar animal was put into another corner. During five minutes we measured the time spent near each of the conspecifics and in neutral area. For both high- and low-anxiety groups, the time spent in the neutral area was less than 60 sec. i.e. rats in a novel environment spent much more time in social contacts than in environmental exploration. Rats with high anxiety spent 88 +/- 32 s and 155 +/- 35 s close to an unfamiliar subject and a cage mate, respectively. On the contrary, the measurements for rats with low anxiety were 200 +/- 40 s for an unfamiliar subject, and 65 +/- 32 s for a cage mate. Consequently, high-anxiety rats preferred contacts with a familiar conspecific, whereas low-anxiety rats preferred to contact with an unfamiliar conspecific.