Teaching hooded crows (Corvus cornix L.) the abstract rule of choice by the matching/oddity of the sample

Six hooded crows (Corvus cornix) were trained in alternative simultaneous matching-to-sample and oddity-from-sample tasks with visual stimuli. Bird's ability to acquire the abstract same/different concept was investigated. For this purpose three stimulus sets of the following categories were used: color (black and white cards), shape (Arabic numerals "1" and "2"), and numerousness (heterogeneous arrays of 1 or 2 elements). These three sets were used for training successively and repeatedly; each time the stimulus set was changed for the next one after the criterion (80% correct or better over 30 successive trials) was reached with the previous one. The training procedure was repeated until the criterion could be reached for each of the three stimulus sets within the initial 30-50 trials. After that, the abstractness of the rule acquired by birds was tested. First, the "partial" transfer tests were conducted, which consisted in presentation of novel stimuli with familiar ones (numerals and arrays in the range from 1 to 4). At the final stage, only sets of novel numerals and arrays in the range from 5 to 8 were presented. The crows demonstrated successful transfer in all of these tests, including the test with absolutely novel stimuli (in the first 24 presentations: min 75.0%, p < 0.02; max 83.3%, p < 0.01). The level of correct choices did not differ from the baseline (p > 0.05). This allows us to conclude that birds are capable for formation of the abstract same/different concept.