RATE OF FORGETTING.--The rate of forgetting is found to be very much more rapid immediately following the learning than after a longer time has elapsed. This is to say that of what one is going to forget of matter committed to memory approximately one-half will fall away within the first twenty-four hours and three-fourths within the first three days. Since it is always economy to fix afresh matter that is fading out before it has been wholly forgotten, it will manifestly pay to review important memory material within the first day or two after it has once been memorized.
DIVIDED PRACTICE.--If to commit a certain piece of material we must go over it, say, ten different times, the results are found to be much better when the entire number of repetitions are not had in immediate succession, but with reasonable intervals between. This is due, no doubt, to the well-known fact that associations tend to take form and grow more secure even after we have ceased to think specifically of the matter in hand. The intervals allow time for the associations to form their connections. It is in this sense that James says we "learn to swim during the winter and to skate during the summer."
FORCING THE MEMORY TO ACT.--In committing matter by reading it, the memory should be forced into activity just as fast as it is able to carry part of the material. If, after reading a poem over once, parts of it can be repeated without reference to the text, the memory should be compelled to reproduce these parts. So with all other material. Re-reading should be applied only at such points as the memory has not yet grasped.
NOT A MEMORY, BUT MEMORIES.--Professor James has emphasized the fact, which has often been demonstrated by experimental tests, that we do not possess a memory, but a collection of memories. Our memory may be very good in one line and poor in another. Nor can we "train our memory" in the sense of practicing it in one line and having the improvement extend equally to other lines. Committing poetry may have little or no effect in strengthening the memory for historical or scientific data. In general, the memory must be trained in the specific lines in which it is to excel. General training will not serve except as it may lead to better modes of learning what is to be memorized.