||It is generally accepted that many vertebrate species possess the ability to formulate concepts. However, using delayed matching-to-sample and oddity-from-sample tasks, Giurfa, Zhang, Jenett, Menzel, and Srinivasan (2001) showed that honeybees may also be capable of utilizing a higher-order cognitive capacity of concept learning. The present study examined the generality of their controversial findings, and was designed to eliminate the notion of perseveration as a reasonable explanation of positive transfer. Here, bumblebees were exposed to a same/different (S/D) discrimination task in which stimulus pairs were jointly presented. It was predicted that extensive training of an S/D relation with one stimulus set would transfer to a novel stimulus set. Only those bees trained with color stimuli and tested with form stimuli demonstrated learning during the training phase, yet bees from both colonies showed immediate transfer to the novel stimuli. In spite of this complexity, current findings lend support to a cognitive interpretation of S/D discrimination in bumblebees.