This ability to recall appropriate facts in given situations is dependent primarily on three factors: power of retention, number of associations, organization of associations. The first factor, power of retention, is the most fundamental and to some extent limits the usefulness of the other two. It is determined by the character of the neurones and varies with different brains. Neurones which are easily impressed and retain their impression simply because they are so made are the gift of nature and the corner stone of a good memory. This retention power is but little, if at all, affected by practice. It is a primary quality of the nervous system, present or absent to the degree determined by each individual's original nature. Hence memory as a whole cannot be unproved, although the absence of certain conditions may mean that it is not being used up to its maximum capacity. Change in these conditions, then, will enable a person to make use of all the native retentiveness his nervous system has. One of the most important of these conditions is good health. To the extent that good blood, sleep, exercise, etc., put the nervous system in better tone, to that extent the retentive power present is put in better working order. Every one knows how lack of sleep and illness is often accompanied by loss in memory. Repetition, attention, interest, vividness of impression, all appeal primarily to this so-called "brute memory," or retentive power. Pleasurable results seem not to be quite so important, and repetition to be more so when the connections are between mental states instead of between mental states and motor responses. An emphasis on, or an improvement in, the use of any one of these factors may call into play to a greater extent than before the native retentive power of a given child. The power to recall a fact or an event depends not only upon this quality of retentiveness, but also upon the number of other facts or events connected with it. Each one of these connections serves as an avenue of approach, a clew by means of which the recall may operate. Any single blockade therefore may not hinder the recall, provided there are many associates. This is true, no matter how strong the retentive power may be. It is doubly important if the retentive power is weak. Suppose a given fact to be held rather weakly because of comparatively poor retentive power, then the operation of one chain of associates may not be energetic enough to recall it. But if this same fact may be approached from several different angles by means of several chains of associations, the combined power of the activity in the several neurone chains will likely be enough to lift it above the threshold of recall. Other things being equal, the likelihood that a needed fact will be recalled is in proportion to the number of its associations.