The design of energy efficient buildings and the potential for using solar energy for heating and cooling is contingent upon optimizing the building ventilation systems. In this paper, we study the ventilation of two interconnected spaces, such as adjacent offices or areas in an open plan office. The goal is to locate return vents to increase the efficiency of night ventilation and to reduce energy consumption. The flow in two interconnected rooms of similar sizes is studied experimentally using a tank divided by an interior vertical wall. A forced buoyancy source with a finite volume flux is located in the ceiling of one-room and an unforced vent is opened in the ceiling of the other room. The goal of the study is to understand the transient cooling/heating that occurs in this two-room system when a forced cold-air vent is located in the ceiling of the first room and a return ventilation exit is located in the second. In particular, we investigate the effects of varying the number of openings and their vertical positions in the interconnecting wall. First, a single opening at the bottom, middle or top of the shared wall is examined. Second, the case of two openings in the wall is considered, with the openings located at the top–bottom, top–middle, bottom–middle, and finally at two mid locations in the wall. The results are compared with the one-room case, which represents the reference case. It was found that, irrespective of the number and locations of the openings, the flow evolves into a quasi-stationary stably stratified two-layer system, with the depths of the layers being different in each room. The average temperature inside each room initially decreases linearly with time and approaches the supply-air temperature at large times. This initial linear decrease holds until cold-air leaves the unforced room through the top-vent at time t e. Subsequently, temperature decreases as an exponential function of time with a characteristic filling time τ = V/ Q s, where V is the total volume of both rooms and Q s is the source volume flux. The efficiency of the ventilation depends on the time t e, and this depends, in turn, on an exchange flow that is established between the two-rooms by the differences in density in each room. For a single opening, the exchange flow takes place as a two-way flow in the opening, while for two openings the flow is from the forced room through the lower opening and in the opposite direction through the upper opening. When the upper opening is located below the ceiling, this flow from the unforced room ‘shields’ the return vent from the dense fluid, thereby increasing the efficiency of the ventilation.