One consequence of Archimedes'' principle is that the mass which a balloon supports in a fluid is independent of the shape of the balloon and depends only upon its volume. For air-filled floating balloons in water some deviations from this result have been observed in first-stage experiments. The dependence of mass on the shape of the balloon which is supported in water has been clearly observed in various observations. It is evident from the first-stage experiments that for floating balloons the principle is only true for particular shapes. The deviations from the principle and contradictions can be explained if the definition of the principle is empirically modified i.e. it is assumed that the upthrust experienced by body is proportional to the weight of the fluid displaced. The constant of proportionality also accounts for shape of body and other relevant factors that were not accounted for by the principle.