A series of low-gravity experiments were conducted to investigate the combustion of supported thermoplastic polymer spheres under varying ambient conditions. The three types of thermoplastic investigated were polymethylmethacrylate (PMMA), polypropylene (PP), and polystyrene (PS). The low-gravity environment was achieved by performing the experiments aboard the NASA DC-9 and the KC-135 Reduced Gravity Aircraft. Spheres with diameters ranging from 2 mm to 6.35 mm were tested yielding Grashof numbers calculated to be less than 0.1. The polymer sphere was supported using a 75-mu m-diameter Al/Cr/Fe alloy wire. The total initial pressure varied from 0.05 MPa to 0.15 MPa whereas the ambient oxygen concentration varied from 19% to 30% (by volume). The ignition system consisted of a pair of retractable energized coils. Two CCD cameras recorded the burning histories of the spheres. The video sequences revealed a number of dynamic events including bubbling, and sputtering as well as soot shell formation and break-up during combustion of the spheres at reduced gravity. The ejection of combusting material from the burning spheres represents a fire hazard that must be considered at reduced gravity. The ejection was found to be sensitive to polymer type, but independent of oxygen concentration and pressure. The average value of the ejection frequency was found to be 3 Hz, 5 Hz, and 5 Hz for PMMA, PS, and PP, respectively. The velocities of the ejected material were estimated by tracking the material in two consecutive video frames. For the PP spheres, V-a = 2.3 (+/- 1.2) cm/s (with 60 events observed). The ejected material appeared to decelerate at an average rate of approximate to 40 cm/s(2), and traverse an average distance of only 8 mm before burning to completion. The V-a for PS and PMMA was not determined because the ejected material was never observed to exist beyond the visible flame of the parent sphere. The average mass burning rates were measured to increase with initial sphere diameter and oxygen concentration, whereas the initial pressure had little effect. The three thermoplastic types exhibited different burning characteristics. For the same initial conditions, the mass burning rate of PP was slower than PMMA, whereas the mass burning rate of PS was comparable to PMMA. The transient diameter of the burning thermoplastic exhibited two distinct periods: an initial period (enduring approximately half of the total burn duration) when the diameter remained approximately constant, and a final period when the square of the diameter linearly decreased with time. A simple homogeneous two-phase model was developed to understand the changing diameter of the burning sphere. Its value is based on a competition between diameter reduction due to mass loss from burning and sputtering, and diameter expansion due to the processes of swelling (density decrease with heating) and bubble growth. The model relies on empirical parameters for input, such as the burning rate and the duration of the initial and final burning periods.