Abstract In a T-maze designed for measuring environmental preferences, restrictedly-fed pigs were offered the choice between a compartment with food and one without food. The food was offered randomly at both sides. In the first series of trials with 5 pigs, all showed position instead of food preferences, in spite of a significant orientation of the head to the food before the choice. In the next series, immediately after a trial the pig was forced to re-enter the starting box and to follow the same procedure again, but this time the gate of the previously visited compartment was closed. By using this “equilibration procedure”, the pigs no longer showed position preferences, but they still did not show a preference for food. Because in both series the pigs were restless while they waited in the starting box before they could make a choice, the waiting times were drastically reduced in the third series of trials. The pigs then behaved more quietly and showed the expected food preference.