Question: The effective volume for level 2 is bigger than that "random coincidence volume" already at log E/GeV = 10.5. Why is this not reflected in fig 6 and for the thresholds given in the text?

Answer: The calculated effective volume for Level 2 is above the "random coincidence volume" at log E/GeV = 10.5, but not with statistical significance. Looking at the error bars provided in fig. 5, the calculated effective volume is still compatible with random coincidences. Thus, only the effective volumes that are statistically incompatible with random coincidences, as noted in the text of the article, are used for fig. 6.