A major defect of all "box" models is the basic assumption that all particles gain and loose energy at exactly the same rate. It is clear physically that there are very large fluctuations in the amount of time particles spend in the upstream and downstream regions between shock crossings, and thus correspondingly large fluctuations in the amount of energy lost. The effect of these variations will be to smear out the artificially sharp pile-ups predicted by the simple "box" models. However our results are based simply on the scaling with energy of the various gain and loss processes together with the size of the acceleration region. Thus they should be relatively robust and we expect that even if there is no sharp spike, the spectrum will show local enhancements over what it would have been in the absence of the synchrotron or IC losses in those cases where our criterion is satisfied.
© European Southern Observatory (ESO) 1999
Online publication: June 18, 1999