Do we know the cosmic-ray boosters?

The literature on cosmic-ray acceleration tends to favour strong shocks (in supernovae, stellar and galactic winds) over compact boosters. Difficulties are posed by the spectrum which does not reveal the implied upper cutoff (at E = 10^13±1 eV), by the energy-dependent anisotropy at the high-energy end (at E 10¹⁸ eV), the missing hydrogen, the large required boosting efficiency, the high positron abundance at E 10 GeV, detailed studies of supernova shells, and by an inadequacy of the Sedov-Taylor model for supernova explosions. Solar and solar-wind data support single-step or few-step accelerations because of the large observed anisotropies and short time scales.

The importance of spindown in compact binary systems is often underestimated. Several neutron-star binaries have been detected at cosmic γ-ray energies. Cosmic-ray escape from the Galaxy may proceed supersonically through narrow tubes rather than via convective diffusion.