| Abstract: | Rats were tested on an elevated radial maze for their ability to choose each of 17 different arms once without repeating any choices. The first experiment indicated that the animals performed well, choosing an average of more than 14 different arms in the first 17 choices. Subsequent experiments demonstrated that: (a) response patterns, general algorithms, or intramaze markings were not necessary for correct choices: (b) there was interference among choices within a test so that the probability of a correct response decreased as the number of choices increased; (c) there was no serial-order effect (primacy or recency); (d) animals tested in a procedure which did not require prior shaping showed no evidence of a general predisposition not to repeat choices. The results are discussed in terms of capacity, accuracy, and other characteristics of working spatial memory. |
