Are Trojan collisional families a source for short--period comets?

We have investigated the dynamical evolution of fragments generated by the impact breakup of Trojan asteroids, in order to study the formation of collisional families in the Trojan clouds and to identify a possible genetic connection between these fragments and the population of short-period comets. Using a collisional model based on the results of high-velocity laboratory impact experiments and previously applied to main-belt asteroid families, we show that typical family-forming Trojan collisions eject a significant percentage (at least =~20%) of the resulting fragments onto unstable orbits. Numerical integrations of the orbits of such fictitious, unstable Trojan fragments show that soon they experience close encounters with Jupiter, and as a consequence they become undistinguishable from those of Jupiter-family comets (some of which are currently close to the 1:1 Jovian resonance), of comets undergoing temporary satellite captures by Jupiter (such as P/Shoemaker-Levy 9), and of objects with Jupiter-crossing or -approaching orbits (such as 944 Hidalgo, 2060 Chiron, 5145 Pholus and 3552 Don Quixote). A reliable assessment of the efficiency of the transfer process from the Trojan clouds into comet-type orbits and of its time dependence is difficult, owing in particular to the poorly known size distribution of Trojans. However, an order-of-magnitude estimate suggests that if the fragment flux from the Trojan clouds over the last =~10^6^yr has been close to the average one over the solar system's lifetime, then a few tens of the 160 known short-period comets might have been generated by Trojan collisions.