The layout of the maze used by Tolman and Honzik (1930a) in a paper about "insight" in rats. If rats are allowed to move from the start box to the goal box by whichever route they choose then the three paths are all tried out, but with experience preferences are established according to length. If path 1 is blocked at A then rats go back and chose pathe 2, which could be a simple example of this preference. If however path 1 is blocked at B, then 9 out of 10 rats in this experiment went back and chose path 3. This is the longest, but the only effective way of avoiding the block at B. Tolman used this sort of evidence in favour of his suggestion that rats form "cognitive maps". A reproduction of the original figure is given in Lieberman (1993/2000; page 489/353).
[on page 19 of the handout]