Abstract
Four rock squirrels received 80 discrimination learning-set (DLS) problems followed by 80 reversal learning-set (RLS) problems in a modified WGTA. Each DLS problem was learned to a criterion of 9 out of 10 correct while each RLS problem consisted of either 4, 8, 12, or 16 pre-reversal trials and 12 post-reversal trials. Significant interproblem learning occurred during the DLS problems. DLS learning suggested that, for squirrels, a criterial DLS procedure is more efficient than presentation of a small fixed number of trials per problem. RLS performance became independent of the number of pre-reversal trials during the last block of 16 problems.

This publication has 11 references indexed in Scilit: