The liability to error resulting from mere oblivescence and the arbitrary selection of mental images is seen most plainly, perhaps, in our subsequent representation and estimate of whole periods of early life. Our idea of any stage of our past history, as early childhood, or school days, is built up out of a few fragmentary intellectual relics which cannot be certainly known to answer to the most important and predominant experiences of the time. When, for example, we try to decide whether our school days were our happiest days, as is so often alleged, it is obvious that we are liable to fall into illusion through the inadequacy of memory to preserve characteristic or typical features, and none but these. We cannot easily recall the ordinary every-day level of feeling of a distant period of life, but rather think of exceptional moments of rejoicing or depression. The ordinary man's idea of the emotional experience of his school days is probably built up out of a few scrappy recollections of extraordinary and exciting events, such as unexpected holidays, success in the winning of prizes, famous "rows" with the masters, and so on. Besides the impossibility of getting at the average and prevailing mental tone of a distant section of life, there is a special difficulty in determining the degree of happiness of the past, arising from the fact that our memory for pleasures and for pains may not be equally good. Most people, perhaps, can recall the enjoyments of the past much more vividly than the sufferings. On the other hand, there seem to be some who find the retention of the latter the easier of the two. This fact should not be forgotten in reading the narrative of early hardships which some recent autobiographies have given us. Not only does our idea of the past become inexact by the mere decay and disappearance of essential features, it becomes positively incorrect through the gradual incorporation of elements that do not properly belong to it. Sometimes it is easy to see how these extraneous ideas get imported into our mental representation of a past event. Suppose, for example, that a man has lost a valuable scarf-pin. His wife suggests that a particular servant, whose reputation does not stand too high, has stolen it. When he afterwards recalls the loss, the chances are that he will confuse the fact with the conjecture attached to it, and say he remembers that this particular servant did steal the pin. Thus, the past activity of imagination serves to corrupt and partially falsify recollections that have a genuine basis of fact.