“The Effect of Floor Cues Upon the Mastery of the Unit-Alike Maze § Curtis1931”, Quin Fischer Curtis1936 (maze-running)⁠:

…Slight slopes and sags in the floor which might have acted as cues were radically altered by elevating the center of the maze with building jacks. The rats were not thrown off…Attempts were made to find a flooring which would be of such uniform character that no differential sounds would be given off in different parts. Among other materials tried were rubberized nainsook, hair felt, cotton batting, 1⁄16-inch lead strips, and Masonite “pressed wood”. Sooner or later, the animals were able to get a differential cue from each of these, and to run any number of like-units without error. This evidence as to the nature of the floor-cue, together with the failure to find any floor which would be uniform for the rats, was the background of the present experiment.

…Curtis1931 attempted to bring the floor cue under control, and to study the learning of the unit-alike maze when kinaesthesis was the only cue allowed the rat. He used a unit-alike maze similar in plan to that used in the present study (see pg26a). The floor of this maze was composed of sand, 1.5-inches deep, over which was spread black oil-cloth. 6 rats were given a long period of training on two standard units, followed by one exceptional unit. When this 3-unit combination was mastered to the extent of the rats’ ability, a series of test runs was given, in which the rats were dropped into the maze at points other than their regular starting point. Results showed that the rats tended to run off a kinesthetic pattern of 3 units, no matter where started in the maze, rather than to pick out the exceptional unit by means of some local cue. Curtis concluded therefore that the rats were not receiving any cues from the floor or elsewhere which would distinguish the exceptional unit from the others in the series.

However, the result of this experiment was not as clearcut as could be desired, because the rats had never mastered the 3-unit pattern with great accuracy, even after long training. A second experiment was therefore performed, in which the rats were trained to run a pattern of only two units, one standard unit followed by the exceptional unit. This combination was readily learned to a high performance level. Then a test series was given as before, with the rats being dropped into the maze at points preceding and following the entrance point used during training. Now, even more certainly than before, the rats ran off a pattern of two units no matter where started in the maze, and without apparent regard for the old exceptional unit to which they had formerly reacted.

Curtis concluded from these experiments that:

1. “It is possible to construct a maze which will afford the rat no environmental cues, and in which his learning may be shown by positive test to be kinesthetic in character.”
2. “The control of floor cues may be effected by use of a floor composed of a deep layer of sand, covered with some uniform material such as oilcloth.”
3. “On a simple unit-alike maze, rats can learn to give a different reaction to one of the units on the basis of its ordinal position in the series, provided that not more than one standard unit precedes the exceptional one.” (pg68)