Light fragments with mass number less than 60 were found to violate limiting fragmentation, and had excitation functions that were strongly increasing with projectile energy until 8.0 to 12.0 Ge-V.With the 1.0 GeV 12 C beam the pattern of mass yields was quite different from that of all the other reactions, with the normal peak in the fission mass region (80 < A < 145), but with much lower yields below mass number 60 and between mass numbers 145 and 210, indicating that these fragments are formed primarily in very energetic reactions in which large excitation energies are transferred to and significant amounts of mass are removed from the target nucleus.Theoretical predictions of the intra-riuclear cascade, nuclear fireball, and nuclear firestreak models are compared with the experimental results. The fireball model is found to be inferior to the other two, due to its failure to deposit large enough excitation energies within the fragment precursors. The intra-nuclear cascade and nuclear firestreak models are both able to predict the general shapes of the experimental distributions, with the exception of the yields for the lightest fragments. However, these two models are found to be incapable of reproducing the typical target fragment recoil velocities, which suggests that some unexpected mechanism must exist for the transfer of large excitation energies to the fragment precursors without the correspondingly large amounts of recoil momenta.