When a hot body A is thermally connected to a cold body B, the textbook knowledge is that heat flows from A to B. Here, we describe the opposite case in which heat flows from a colder but constantly heated body B to a hotter but constantly cooled body A through a two-phase liquid-vapor system. Specifically, we provide experimental evidence that heat flows through liquid and vapor phases of cryogenic helium from the constantly heated, but cooler, bottom plate of a Rayleigh-Bénard convection cell to its hotter, but constantly cooled, top plate. The bottom plate is heated uniformly, and the top plate is cooled by heat exchange with liquid helium maintained at 4.2 K. Additionally, for certain experimental conditions, a rain of helium droplets is detected by small sensors placed in the cell at about one-half of its height.O ur cryogenic experiment takes place in a cryostat containing a Rayleigh-Bénard (RB) convection cell shown in Fig. 1. The cylindrical RB cell, 300 mm in both diameter and height, is designed to minimize the influence of its structure on the convective flow (1) and is capable of running at very high Rayleigh numbers up to 10 15 (2, 3). These studies used cryogenic helium gas as the working fluid and have been performed under nearly Oberbeck-Boussinesq conditions (all physical properties of working fluid assumed constant except its density varying linearly with temperature) (2), as well as for the case when non-OberbeckBoussinesq conditions cause asymmetry between the top and bottom boundary layers (3). Here, we report results of experiments on two-phase heat transport, using cryogenic helium vapor and normal liquid 4 He as working fluids with remarkable, well-known and in situ tunable properties (4, 5). We stress that the heat conductivity of the thin stainless-steel cylindrical wall and any possible parasitic heat input to the RB cell are negligibly small in this work.We start very near equilibrium conditions, with the RB cell filled one-half with normal liquid helium and one-half with helium vapor. The temperature of the cell is approximately that of the thermally connected liquid helium vessel (LHeV), as shown for time t < 0 in Fig. 2A, before the homogeneously distributed resistive heater in the bottom plate is turned on at t = 0. This condition closely corresponds to a point on the equilibrium saturated vapor curve (SVC), calculated based on the continuously monitored value of pressure, P, in the cell. Then we start heating the bottom plate with a constant input power into the resistive heater, and, using built-in germanium thermometers (1, 2), continuously monitor the temperatures T B and T T , of the highly conductive bottom and top copper plates, respectively, as well as the temperature readings T 1 . . . T 4 of four small Ge sensors (6) installed within the cell as shown in Fig. 1. The temperature of the upper plate is not controlled; it is only affected by heat exchange with the helium inside the cell and by the weak thermal link to the LHeV. For an isolated system consisting of liquid and...