The time interval distributions of the Barkhausen impulses were investigated. The experimental results show beautiful regularities, in contrast with Tyndall's conclusion. An explanation for the experimental result was given, by assuming a simple mechanism for the reversal of domain magnetization. It was concluded that the magnetic reversal is not perfectly accidental, as is the case in the disintegration of radioactive substance. The reversal of domain magnetization needs always some accumulation of energy by some effects, either magnetic or of origins of other kinds.