The rhythmical and other _groups_ that are found or made by the learner in memorizing nonsense lists are, in effect, "higher units", and have much the same value as the higher units of telegraphy or typewriting. One who learns many lists in the course of a laboratory experiment develops a {335} regular system of grouping. First he reads the list through, in groups of two, three or four items, noticing each group as a whole; later, he notices the items in each group and how they are related to each other. He also notices the interrelations of different groups, and the position of each group in the total series. All this is quite different from a mere droning along through the items of the list; it is much more active, and much more observant. Very interesting are the various ways in which the learner attacks a list of nonsense syllables, numbers, or disconnected words. He goes to work something like the cat trying to escape from a strange cage. He proceeds by a sort of trial and error observation; he keeps looking for something about the list that will help to fix it. He sees something that promises well for a moment, then gives it up because he sees something better. He notices positions, i.e., connects items with their position in the list. He finds syllables that stand out as peculiar in some way, being "odd", "fuzzy", smooth, agreeable, disagreeable, or resembling some word, abbreviation or nickname. He notes resemblances and contrasts between different syllables. He also finds groups that resemble each other, or that resemble words. Besides what he actually finds in the list, he imports _meanings_, more or less far-fetched, into the list. He may make a rhythmical line of verse out of it; he may make a story out of it. In short, he both explores the list as it stands and manipulates it into some shape that promises to be rememberable. His line of attack differs according to the particular test that is later to be made of his memory. Suppose he is shown a number of pictures, with the understanding that later those now shown are to be mixed with others, and that he must then pick out those now shown--then he simply examines each picture for something characteristic. But {336} suppose each picture is given a name, and he must later tell the name of each--then he seeks for something in the picture that can be made to suggest its name. Or suppose, once more, that the pictures are spread out before him in a row, and he is told that they will later be mixed and he be required to rearrange them in the same order in which they are now shown--then he seeks for relationships between the several pictures. His process of memorizing, always observant, exploratory and manipulatory, differs in detail according to the memory task that he expects later to perform. For another example, suppose an experiment is conducted by the method of "paired associates". The subject is handed a list of pairs of words, such as