The Viewpoint by Bruce Winstein and Kathryn M. Zurek [1] highlights the recent measurement of high-energy cosmic-ray electrons by the Fermi Gamma-Ray Space Telescope [2], in the context of two other recent electron measurements—the PAMELA spacecraft [3], which operates at lower energy than Fermi but distinguishes electrons from positrons, and the ATIC balloon-borne instrument [4], which presents results in the same energy interval as Fermi and like Fermi does not distinguish electrons from positrons. Their commentary is valuable for its description of the astrophysical significance of these cosmic-ray electron results.

However, in my view Winstein and Zurek are in error when they refer to the distinct difference between the Fermi and ATIC results and state unequivocally “… the large ATIC excess … is ruled out.” To be sure, the Fermi results have better statistical precision than the ATIC results, but the difficult part of this high-energy electron measurement is in the systematic uncertainties—distinguishing electrons from protons, which are ~10⁴ times more abundant at these energies, and accurately measuring electron energy in a steeply falling energy spectrum (flux proportional to E^-n with n~3). Both the ATIC and the Fermi scientists have carefully examined their systematic uncertainties and both have a great deal of confidence in their results. Of course, the discrepancy between the two makes clear that one or the other (if not both) contains some error, but we cannot be certain right now where the error lies.

Neither the Fermi telescope nor the ATIC instrument was designed primarily for the electron measurements reported here. The Fermi telescope was optimized for gamma rays up to 300 GeV; it does not have the depth of the ATIC instrument for fully containing the higher-energy electron showers. The ATIC instrument was optimized for high-energy cosmic-ray nuclei; it does not have the fine spatial resolution for visualizing the beginning of the shower that the Fermi telescope has.

It is clear to me, and I believe it is also clear to both sets of scientists, that both groups have work to do to understand and resolve their differences. The Fermi results may turn out to be correct, but in my view it is premature to declare unequivocally that Fermi is right and ATIC is wrong.

Winstein and Zurek reply: We agree with Martin Israel that, to resolve the discrepancy, both experiments need to work to be sure that the systematic uncertainties are correct. Still, we conclude that the evidence for a prominent feature in the spectrum has been cast in serious doubt. And since our Viewpoint appeared, the HESS team has released a measurement of the e⁺e^- flux in the energy range 400 GeV – 5 TeV [1], overlapping with the Fermi result. HESS and Fermi agree (on the lack of a feature), each with very high statistical significance. It is possible that both HESS and Fermi have treated their systematics incorrectly, causing each to miss the feature observed by ATIC, but the evidence at this point is leaning in the direction of Fermi and now HESS: there is no prominent feature in the 400–800 GeV range. As Israel points out, the most crucial systematic uncertainties involve the behavior of the calorimeters, and those of Fermi and ATIC have relative strengths and weaknesses. Experiments with high statistics are generally better able to probe unforeseen systematic uncertainties than are those with low statistics. We, along with Israel, look forward to having these discrepancies resolved by the experiments in question.