A new method for estimating asteroid collision probabilities and impac
t velocities (Vedder 1996, Icarus, 123, 436-449) is extended from sing
le to multiple targets. It is then applied to estimating collision rat
es and velocities of all asteroids in the main belt with each other. T
he results obtained are compared with those obtained by previous inves
tigators using other methods. The new method yields main belt collisio
n probabilities about 10-15% higher than earlier results, and the impa
ct velocity distribution is shifted downward by about 20%. Analysis of
the velocity components shows that the most significant difference is
in the component normal to the plane of the ecliptic, with the new me
thod yielding a higher likelihood of this Z component having a low mag
nitude. The nonnormal distribution of the Z component in previous resu
lts may be due to underestimating the collision probability for nearly
coplanar or nearly tangent orbits. For these frequently occurring cas
es, the new method may well have advantages, because it is singularity
-free and does not depend on an elaborate and very sensitive analytica
l model of the geometry of close encounters. (C) 1998 Academic Press.