“…What makes this channel particularly apt for an investigation of the cusp, apart from the enormous statistics collected by the NA48/2 collaboration [3], is the significantly larger branching fraction of K + → π + π + π − compared to K + → π 0 π 0 π + , such that the perturbation of the decay spectrum of the latter is very sizeable. This is in marked contrast to two other decays that have been studied subsequently and that display, in principle, cusp structures generated by the same mechanism: K L → 3π 0 [2,4,8,9], and η → 3π 0 [8,[10][11][12][13][14]. In both of these, the weaker coupling to the charged-pion final state diminishes the cusp to a mere 1-2% effect on the decay spectrum.…”