The cosmic ray spectrum at ultra high energies (E > 1 EeV) has two features: the ankle near ~3 EeV and the so-called second break point, at ~60 EeV. If cosmic rays were pure protons at the highest energies, the second break point is explained by the well known Greissen-Zatsepin-Kuz'min process: energy loss of the protons on cosmic microwave background (CMB) due to the photo-pion production. In the case of a mixed chemical composition, the prediction for the position of the second break point is complicated by the spallation of the heavier nuclei on the cosmic microwave background. In the second case, the energy of the second break point could be lower than that of pure protons. Both Telescope Array and Pierre Auger experiments have measured the spectrum at the highest energies, and the results are in good agreement from 0.1 to ~25 EeV when the two measurements are adjusted to use a common energy scale. Above ~25 EeV, however, there is a significant discrepancy between the two results: the second break point in Pierre Auger spectrum occurs at a significantly lower energy than that of the Telescope Array. This effect cannot be explained by adjusting the energy scales of the two experiments.
In this work, we use data of the Telescope Array surface detector to show evidence of the dependence of the second break point energy on the declination. When we restrict the TA declination to a range from -15 to 24.8$^{\mathrm{o}}$, we see that the second break point occurs at a lower energy of ~40 EeV, in better agreement with the Pierre Auger result. The difference between the TA low and high declination break points is a 3.9$\sigma$ effect. Also, we perform checks of the systematic uncertainties and demonstrate that this is not an instrumental effect.