Limited capacity information processing and pigeon matching-to-sample: Testing alternative hypotheses