Consideration of experimental studies on animals and humans indicates that a number of different phenomena are ordinarily included in the general concept of "anxiety." The varieties of precipitating events and the varieties of responses are sufficiently great that these phenomena are not useful as criteria of a single, general class of events. There is one phenomenon, however, which does differentiate a class of events that would be useful to identify whether or not it is called "anxiety." This phenomenon is change in the form of behavior (i.e. the mode of the interaction pattern) which occurs in the course of a transaction. Such a change is always present in situations with humans in which "anxiety" is diagnosed, but may also occur in situations in which the verbal, gestural, and autonomic manifestations are not characteristic of clinical anxiety.Study of this phenomenon obviously requires more or less continuous observations over the temporal course of a transaction, permitting examination of the precipitating events and later developments. For this purpose animal experiments are preferable since they are simpler and are not contaminated by post hoc verbal rationalizations.In general the precipitating events have one or other of 3 characteristics. These are: (1) the more or less sudden arrival of information of environmental response inconsistent with the information the anticipatory behavior is, as it were, prepared for; (2) arrival of information previously associated with situations inevitably (i.e. regardless of any performance the animal is capable of) including "pain" or "doom"; and (3) change in rate or in variability of information such that the interacting system becomes overloaded and functionally disorganized.The initial responses to the precipitating events would appear to include increased activity in one or another of the mesencephalic and diencephalic mechanisms which either facilitate alertness and capacity for rapid change in anticipatory "set," or which provide somatic, particularly visceral, support for patterns of interaction requiring high energy output, such as attack, flight, etc. It is now quite clear that there is a reciprocal interaction between these "alarm" and "emergency" systems and the anticipatory, information processing mechanisms of the forebrain. The anatomical substrata are in part included in the reticulo-thalamo-cortical ascending system of tracts and in the descending cortical connections to the tectum, the reticular formation, and the subthalamic centers, as well as the descending tracts from the limbic lobe. Activity in certain of the last named is apparently capable of inhibiting parts of the alerting and emergency-action functions. It may be noted here that the variety of innate patterns of behavior mediated by the brainstem is such that the general concept of a diffuse mesencephalicposterior hypothalamic "emergency," "alerting" or "activating" mechanism must be recognized as only a first approximation. Indeed, these aspects of the activity of the systems involved may be secondary functions supportive to relatively precisely integrated behavior patterns.The later behavior following the more or less acute change in mode of transaction shows wide range and considerable modifiability. The only common characteristic of the numerous patterns seen would appear to be that they result in decrease in the area of interaction—both spatial and temporal—between the organism and the environment. This appears to be accomplished either by reduction of the motility of the organism or by limitation of the scanning activity of the information processing systems.It is important to note that the careful, quantitative studies of animal behavior in experimentally limited situations and with attention to the time axis permit differentiation classes of behavior not readily apparent in field or clinical observations. In particular, the sharp differentiation of two types of "learning," one abolished and the other virtually unaffected by electroconvulsive stimulation and also by reserpine, represents an important contribution not only to learning theory but also to knowledge of the anxiety-type of phenomena. Concurrent studies of physiological, endocrinological, and psychological phenomena promise further differentiation both of patterns of disturbances of anticipatory behavior and also of the control mechanisms which may be sequentially brought into play.