A TEST OF COMPARISON-STIMULUS SUBSTITUTABILITY FOLLOWING ONE-TO-MANY MATCHING BY PIGEONS

Pigeons were initially trained to match each of two sample stimuli to two different comparison stimuli (one-to-many matching). Afterwards, they learned to match one comparison from each associative set to new samples. Then, in a test phase, the remaining comparisons were substituted for those used in the interim phase. Two birds showed evidence of transfer during testing but the majority matched at accuracy levels close to chance. This relatively weak transfer effect contrasts with more robust transfer obtained after other training regimens (Urcuioli, Zentall, Jackson-Smith, & Steirn, 1989), suggesting a possible ''direction-of-training'' effect in producing emergent stimulus relations in pigeons' matching-to-sample.