We argue that the unexpected nonexponential behavior of some cosmic ra
y data is just a manifestation of cross section fluctuations discussed
recently in the literature and observed in nuclear collisions and in
diffraction dissociation experiments on accelerators.