In the investigation of the German folksongs, one of the most basic problems has always been their vast quantity. Indeed, their sheer number is the essence of this form, which is based on an expression of musical universality and is, therefore, easily understood and immediately accessible.' Specifically, these songs are characterized by the same "natural"2 monodic, strophic, periodic, and tonal-harmonic principles. Furthermore, each song is capable of appearing in numerous versions, according to the individuality of each performer. Perhaps somewhat paradoxically, this tradition has contributed to both the large number and the basic similarity of German folksongs. It is precisely in this inherent structural homogeneity that we find the key to dealing with such a large number of individual phenomena. This homogeneity, whose specific qualities will be discussed shortly, makes the problem of sheer numbers more manageable: the closer the structural relationship of the given body of songs, the fewer the number of attributes needed to distinguish them from one another. Stated quite simply, the more characteristics all songs hold in common, the easier to differentiate between individual songs. The only question is exactly how and in what proportion these songs differ from and resemble one another. Thanks to the computer, the mass manipulation of such a large body of material presents little problem today. The characteristic structural simplicity of the German folksong makes it particularly well suited to the capabilities of this instrument. It is, in fact, puzzling that German folksong research has not already made use of this valuable tool.3 Material The material under consideration in this study consists of German folksongs which have appeared in print since the end of the eighteenth century.4 It is characteristic of these melodies that they (1) are transcribed according to the system of bar-line notation; (2) are based on the principle of corresponding paired measures; and (3) conform to the principles of major-minor tonality. These criteria, fundamental to the German folksong, were deemed particularly important in the collection of material.s The question of whether or not certain melodies lend themselves to bar-line notation6 is thus irrelevant here by virtue of the transcription methods in use since the pre-Classical era. The material outiined above, then, has the homogeneity necessary for mass manipulation by the computer. Code A first step toward thle study of folksong melodies through the computer is the development of a code for entering and storing those melodies, which takes into consideration the nature, range, and goal of the investigation. While attempts have been made toward a universal code encompassing all symbols of musical notation,' such a code would be both unnecessary and impractical for specialized studies such as this one: not only would the input and calculation processes be cumbered by the large number of musical symbols not needed here, but too much space would have to be reserved for the storage of these never-to-be-used symbols.8 At the same time, the code must guarantee as clear and complete a representation of each melody as is necessary for our purposes. To be practical, therefore, the code must (1) be easily understood, i.e., syntactically simple and lucid; (2) be easily learnable, consisting of as few (preferably mnemonic) symbols as possible, to decrease the chances of error; (3) easily manipulable, to keep transcription time to a mini-