For this experiment a form of maze was used which was sufficiently difficult for the human subject and involved only a single type of cul-de-sac. Two patterns of the maze were constructed, the only difference in the two being that one was a lateral reversal of the other. Sixty subjects took part in the experiment. The results support the view, generally held for rote learning, that primacy and recency are important factors in determining the formation of motor responses. The fact that the end positions were mastered first and the middle last weakens the view that the regressive order of elimination is characteristic of maze learning. There was a marked tendency for the right hand to enter more frequently and persistently the culs-de-sac located on the left half of the maze, and for the left hand to enter those on the right half. Culs-de-sac situated far outward to the right or left of the median line were uniformly eliminated early in the series. For some obscure reason the first pattern of the maze proved to be the more difficult of the two. That such a small difference in the maze should alter the results to such an extent suggests the unreliability of making direct comparisons of data obtained from mazes that differ greatly in pattern. From Psych Bulletin 22:05:00411. (PsycINFO Database Record (c) 2006 APA, all rights reserved)