Altruistic behavior in lower animals has been investigated by controlled experiment (Lavery & Foley, 1963; Massennan, Wechkin, & Terris, 1964;Rice & Gainer, 1962), but there are mixed conc1usions about the existence of such behavior and about its detenninants. The research described here was designed to establish more conclusively the presence or absence of altrilistic behavior in the white rat. Rice & Gainer's (1962) operational definition of altruism was used, Le., "behavior of one animal that relieves another animal's distress." A distress state (manifested by squealing, urination, defecation, and a "mincing dance") was induced in a rat (Victim or V) by administration of electric foot shock. Another rat (Operator or 0) could introduce or terminate this distress state by appropriate manipulation of two levers. SUBJECTS The Ss were 70 male albino rats (56 Os and 14 Vs) of a commercial Sprague-Oawley strain, 90 to ISO days of age. APPARATUS Apparatus was an enclosed box divided ihto two adjacent compartments by a glass wall; spaces around the edges of the glass wall allowed passage of visual, auditory, and olfactory stimuli. The compartment for the V had three bare walls, and a grid floor through wh ich electric shock could be delivered to the feet of the V. Constant shock delivery of .05 mA to the grid was assured by use of a commercial electro-mechanical scrambling device which changed the polarity of each grid-bar four times/sec. In the compartment for the 0 were two combination operant lever-food tray devices, each mounted on a wall adjacent to the glass separating wall. Each lever-tray device had a food-pellet dispenser which was set to deliver one pellet of food per lever press throughout the experiment. Circuitry was such that either lever could be set to actuate or terminate the grid shock in addition to and concurrent with the delivery of a food pellet. Thus, it could be arranged that a press of lever X by the 0 would deliver one pellet of food to the 0 and either introduce or terminate shock to the V, while a press of lever Y would yield only food reward, or vice versa. An automatie timer and reset system was used so that, for each lever press, shock was either introduced or tenninated for 3 sec. Ouring all experimental trials the entire apparatus was operated automatically, PROCEOURE Each 0 underwent 12 to 16 30-min trials (training and experimental), conducted on consecutive days. Os were deprived of food for 23 h prior to each trial.Each 0 was first trained to feed itself by operating either lever, and then was allowed to develop a preference for one lever. A lever preference was considered to exist when the 0 procured more than 90% of his food pellets by use of the same lever for two consecutive 30-min trials. To speed up the development of a preference, one lever was adjusted so that it Psychon. Sei., 1969, Vol. 14(1) required twice as much force to activate as did the other. By the end of 8 to 10 training trials over 80% of the Os showed a preference for the "easier" lever; remaining Os preferre...