Abstract
One of 2 groups of 8 cats and of 5 monkeys was trained on 60 reversals of a single discrimination before being trained to criterion on 80 different visual discrimination problems with object cues relevant and position cues irrelvant; for the others positional cues were relevant and object cues irrelevant. Both experimental groups of monkeys made significantly fewer errors on the multiple problem series than 5 control monkeys, tested without prior reversal training. No differential transfer effects were observed among experimentals and 11 control cats. Mammalian species differ greatly with respect to intertask transfer; monkeys trained on repeated reversals seemed to acquire a "win-stay, lose-shift" strategy that facilitated formation of learning sets, and cats did not.

This publication has 5 references indexed in Scilit: