An experimental study is conducted to determine the circulation patterns inside a rectangular enclosure due to natural convection using a Particle Image Velocimeter (PIV). Experiments were conducted using two different fluids (air and water) and for rectangular enclosures with aspect ratios 0.5 and 1.0. Natural convection in enclosures has been experimentally studied in the past. Many of these studies cited in the literature use some kind of an optical method like interferograms, shadowgraphs, streak photographs, or multi-exposure photographs to visualize the flow patterns in the enclosure. The present study employs a commercial two-dimensional PIV to capture, instantaneously, the circulation patterns inside the test section. The test cavity in the present setup is of rectangular shape, which is 5 inches (127 mm) wide, where the height of the enclosure can be changed to obtain aspect ratios of 0.5 and 1.0. The depth of the rectangular enclosure measures 12 inches (305 mm) to minimize the effect of walls normal to the two dimensional flow patterns that are expected in this type of arrangement. The walls of the cavity are made of Aluminum plates. These plates are kept at constant but different temperatures during the experiments. In the present study, hollow glass sphere particles with 10 microns in diameter were used as seeding for water experiments and fine particles/flakes of ash generated from burned incense were used as seeding in the air experiments. For each working fluid, the experiments were repeated for different aspect ratios and for different wall temperature differences which corresponded to Rayleigh numbers in the range of 106 and 107. Velocity fields were captured at steady state for each experiment using the two-dimensional PIV system. Numerical studies were also carried out using a commercial CFD software. Comparisons of the numerical and experimental results indicate a good match in terms of circulation patterns and velocity magnitudes in the core of the buoyancy driven flow. Discrepancies in measured and predicted values of velocities are more pronounced near to the boundaries of the enclosure. Separate measurements with finer interrogation areas and different PIV setting were required to improve the accuracy of the measurements near the corners (top and bottom) of the enclosure. The results of these measurements are also presented.