Aspect ratios of ellipsoidal bubbles in clean and contaminated systems were measured to investigate the effect of surfactant on the bubble aspect ratio. The bubble diameter ranged from 0.80 to 5.8 mm. Air was used for the gas phase and the glycerol-water solutions were used for the liquid phase. The experimental ranges of the Mortor number, M, the bubble Reynolds number, Re, the Eotvos number, Eo, the Weber number, We, and the Tadak number, Ta, were as follows: 10(-7) < M < 10(-3), 3.6 x 10(-1) < Re < 1.7 x 10(2), 1.1 x 10(-1) < Eo < 5.9 5.0 x 10(-3) < We < 3.7 and 5.3 x 10(-2) < Ta < 3.6. Triton X-100 and 1-octanol were used for surfactant The terminal velocities of bubbles were confirmed to be independent of the surfactant concentration, ant therefore, the bubbles were fully-contaminated from the point of view of the terminal velocity. By making use of the aspect ratio data, the applicability of available aspect ratio correlations was examined and an empirical correlation was proposed. The applicability of the proposed correlation was also examined for sodium dodecy sulphate (SDS) and 1-decanol. As a result, the following conclusions were obtained: (1) the aspect ratio database which is of use for validating and developing shape correlations, was obtained, (2) the available aspect ratio correlations are not applicable to the present data for fully-contaminated ellipsoidal bubbles, (3) the aspect ratios are well correlated in terms of a combination of the Mir& and bubble Reynolds numbers, which prove; that all the relevant forces, i.e. the viscous, inertial, buoyant and surface tension forces, should be taken into account to correlate bubble shapes, and (4) the empirical correlation of the aspect ratio developed by using the Triton X-100 and 1-octanol data also gives good evaluations for bubbles contaminated with SDS and 1-decanol it spite of largely different adsorption-desorption characteristics.