Capacity of birds for transitive inference: the solution of the Gillan test by corvids and pigeons.

A system of four to five differentiations forming a series, in which the magnitude of the reinforcement (R) of each succeeding color stimulus was greater than the preceding by one unit, was formed in pigeons (Columba livia L.) and crows (Corvus corone cornix L.). The number of units of the R that was associated with each of the stimuli increased successively from stimulus pair to stimulus pair from one up to four to five (1st series) and from five to ten (2nd series). After the development of the system of differentiations, the capacity of the birds to compare the magnitudes of R in given new combinations of stimuli and to choose the stimulus associated with the greater R (transitive inference) was tested. Given small Rs, the pigeons and crows solved the test equally successfully; in the case of large reinforcements, the pigeons began to make random choices, while in the crows the proportion of correct choices decreased. The proportion of appropriate solutions was greater when it was a greater absolute difference between the number of units of the R to be compared. The solution of the test employed by the birds is not proof of their capacity for transitive inference, since it can be explained by their capacity for unexpected comparison of the absolute values of the R associated with each of the stimuli.