Observational constraints on the Centaur population

Spacewatch was the first large-scale astronomical survey program employing automated and real-time software fur moving object detection. The search has resulted in the discovery of three new Centaurs and a serendipitous detection of (2060) Chiron. We have determined the efficiency of the Spacewatch system as a function of an object's apparent visual magnitude and rate of motion. Convolving these efficiencies with the scanning pattern of the telescope and a theoretically motivated orbital distribution for the Centaurs allows us to determine the system's efficiency as a function of their absolute magnitude. The efficiency is nonzero in the range -4 < H < 11 (objects larger than about 40 Irm diameter) and has a maximum of about 3.4% at H similar to 5 (similar to 600 km diameter). Restricting this study to a pristine set of data left a single Centaur discovery for this analysis. Assuming that the magnitude-frequency relation follows a power-law distribution (proportional to 10(alpha H)), and based on the detection of 1993 HA(2), we find alpha = 0.61(-0.40)(+0.70)(stat) +/- 0.08(sys). At the 99% confidence level, for the orbital distribution used in this analysis and in the absolute magnitude range -4 < H < 10.5, there must be fewer than similar to 2000 Centaurs. This implies that the Centaur population may be as great or larger than the set of Main Belt asteroids in the same size range. Our best estimate suggests that there may be a total of only three objects of Chiron's size (similar to 200 km diameter) or larger. (C) 1997 Academic Press.