The characterization of the female breast by ultrasonic visualization techniques presents unique problems in comparison to visualization of other body areas. Basic to this uniqueness is the fact that, not only does the breast contain a variety of types of tissues, but the amount of each type of tissue is a variable. Correct identification, from the echogram presentation, of abnormalities located within any of these tissue structures is partly dependent on the ability to correctly identify normal echogram patterns over the full range of this variability; such identification necessitates precise correlation of the normal echogram pattern with the specific tissues and with their orientation in the path of the sound beam. However, this type of correlation is difficult to obtain in the in vivo breast. This paper discusses possible source of errors in such attempts. The basic data required to prevent such errors of interpretation of breast echograms can be obtained by carrying out a detailed acoustic visualization study of excised whole breast tissue, followed by correlation of the echogram patterns with the gross and histological sections of the breast tissue. Data obtained to date, using this approach for both normal excised breast tissue and excised tissue containing malignant abnormalities, will be presented.