Very simply put, I have an intense desire to understand an experimental result which, on the surface, violates entropy and the arrow of time-- although, since the experimenters predicted exactly that outcome, a deeper analysis must show that it does not actually violate entropy and the arrow of time-- I’d like to have that deeper analysis explained to me. . The experiment, which can be found here (https://arxiv.org/abs/1711.03323 ), involved first cooling the Carbon portion of a chloroform molecule and warming the Hydrogen portion, and also establishing a quantum correlation between those two atoms. Then, as the correlation decohered, the experimenters observed-- as predicted-- a flow of heat from the cold Carbon to the warm Hydrogen, contrary to what would have occurred, of course, if the two atoms had not been correlated. I realize that the correlation, when it existed, constituted ‘information’, and so the correlation’s decoherence meant information loss and an entropy increase, which in turn meant the release of heat. My question is: why wasn’t all of the heat produced by the decoherence of the correlation simply lost to the environment--why did some of that heat flow from the cold Carbon to the warm Hydrogen causing an entropy decrease there, and thereby a reduction in the net entropy gain for the system overall (the entropy gain from the correlation’s decoherence was diminished by the entropy loss from the heat flow from the cold Carbon to the warm Hydrogen atom-- strikingly odd behavior for heat that would not have occurred if the two atoms had not been correlated to begin with)? Evidently the initial existence of the correlation dictated that very strange cool-to-warm heat flow behavior, but why?