Thermal diodes of convective type - the research object presented here - are solar passive devices destined to building heating. Figure 1 presents the shape and principle of operation of the thermal diode. The diode operation may be divided into two phases: warm-up phase during the solar radiation and cool-down phase at night or during a cold cloudy day. The existing literature about the thermal convective diodes is very scarce; only the work of Jones [1] is available. In Chapter 2 the mathematical description of heat transfer in the diode collector is presented for the warm-up phase. The quasi-steady temperatures of liquid were assumed as well as ideal liquid mixing in an accumulator and constant physical properties of liquid (excluding density). The initial convective layer thickness (delta(o)) were determined experimentally. Some symbols of the model and the system of co-ordinates assumed are presented in Fig. 2. Convective heat transfer in the diode collector was described by the Fourier-Kirchhoff equation (1); the temperature of liquid and velocity distributions are given by Eqs. (2) and (3), respectively. Boundary conditions for temperature and velocity functions were written down in Eqs. (4a-4c) and (5a-5d). Combining equations (1) through (5) the solutions were obtained for local temperature (7), local velocity (10) and convective layer thickness for a general case (Eqs. (12)-(14)) and for the no heat loss case (Nu0 = 0, Eq. (17)). The energy balance for the warm-up and cool-down phases was also presented (in two alternative forms, Eqs. (19), (20)). A number of the laboratory experiments described in Chapter 3 were carried out in order to verify the theoretical model proposed. These experiments included the measurements of the distribution of the thickness of the convective layer along the collector and the distribution of the local temperature of liquid along and across this layer for various heat fluxes. The experimental equipment is presented in Fig. 3. Heating the diode collector with an electric heater equipped with thick insulation allowed the simple Eq. (17) (no heat loss case) to be applied. The experimental results obtained are presented in Figs. 4, 5, 6 (points) vs. theoretical values calculated from Eqs. (7), (10) and (17) (solid lines). These figures show satisfactory agreement between the experiments and theoretical considerations. The experiments were carried out over three-year period of time; thermal diodes set was placed in a testing cell and exposed to various weather conditions. Heat transfer coefficients k(i), k(z) and k(o) were estimated from the results obtained using Eqs. (19) and (20). The relationship between the coefficients k(z), k(o) and the insulation parameters (Eq. (21)) gave the value of k(z,obl) by 20% smaller than that estimated basing on the measurements. This agreement was assumed as satisfactory. It was found that the thermal resistance of insulation between the accumulator and the collector has a considerable effect upon the efficiency of thermal diode.