Energy dependence of cosmic ray composition above 10(15) GeV/nucleus

It is argued that above 10 to the 5th power GeV/nucleus, in the range where charge-resolved spectra have not yet been determined, the appropriate measures of equal-energy composition are 1nA and 1nA , the mean value and dispersion relative to the mean value and dispersion relative to the mean of 1nA, where A is the mass number. Experimental data which are sensitive to changes in 1nA with increasing energy are examined. It is found that, taken as a whole, they show no change (+ or 0.5) between 10 to the 5th power and 10 to the 6th power GeV, and a decrease of 1.5 + or - 0.5 between 10 to the 6th power and 10 to the 8th power GeV, with no further change + or - 0.5) above 10 to the 8th power GeV. Taken as a whole, the various indirect estimates of the absolute value of 1nA above 10 to the 5th power GeV/nucleus are also consistent with this pattern. For a wide range of astrophysically plausible composition models the value of the other measure, 1nA is insensitive to changes in 1nA . Because of this the existing data on 1nA can likewise easily be reconciled with this pattern.