Removal of dead ants from the nest (necrophoric behavior) is released solely by contact chemical cues in the fire ant, Solenopsis invicta Buren. Exhaustively extracted corpses do not release necrophoric behavior, but the extracts do when applied to filter paper bits. The necrophoric releaser is absent at death but appears rapidly and reaches a plateau within about an hour. The rate of signal appearance is identical in heat and freeze killed workers, implying a non-enzymatic origin. There is no specialized caste or size of worker which carries out necrophoric labor. In the field, in the absence of slope, corpse-bearing workers head outward from the nest on random radii and drop their corpses at unpredictable distances, making refuse piles rare. There is a positive relationship between slope and the presence of refuse piles, and these are located downhill from the mound. When the headings of necrophoric workers were measured in circular arenas in the lab, the only potential orientational cue (tested : landmarks, light, 5°, 10°, 15° slope) which resulted in non-random distribution of headings was slope. The concentration of the headings was a direct function of the slope and seemed to plateau at about 15°. Corpse-bearing ants show stereotyped behavior upon encountering refuse piles and adding their burden to it. Chemical stimuli probably issuing from the feces in the refuse pile bring about the end of necrophoric behavior and maintain the refuse pile. These chemical cues, as also those initiating necrophoric behavior, must be contacted to be effective.