The technique of computerized tomography has been used earlier to measure point density (or void-fraction) distribution in air-water flows. The interrogation source used in the measurement system was a gamma ray source, Cs-137. The inherent nature of such sources is the Poisson nature of the emmitted radiation. This Poisson corruption leads to imperfections in the projection data, i.e., the data no longer represents the line integral of the desired property being measured. The present study was conducted to assess the sensitivity of the tomographic reconstruction process to the statistical (Poisson) uncertainity in the projection data. Reconstruction experiments performed on six different sets of data indicate that influence of Poisson errors on reconstruction profiles is much more pronounced near the centre of the pipe in all the cases indicating the need for stronger gamma ray sources than the 13.6 milli-Curie used in the earlier investigations.