[Untitled]

In the atmospheric Cerenkov technique gamma rays are detected against the abundant background produced by hadronic showers. In order to improve the signal to noise ratio of the experiment, it is necessary to reject a significant fraction of hadronic showers. Traditional background rejection methods based on image shape parameters have been extensively used for the data from imaging telescopes. However, non-imaging Cerenkov telescopes have to develop very different means of statistically identifying and removing cosmic ray events. Some of the parameters which could be potentially important for non-imaging arrays are the temporal and spectral differences, the lateral distributions and density fluctuations of Cerenkov photons generated by gamma ray and hadron primaries. Here we study the differences in fluctuations of Cerenkov photon density in the light pool at the observation level from showers initiated by photons and those initiated by protons or heavier nuclei. The database of simulated events for the PACT array has been used to evaluate the efficiency of the new technique. Various types of density fluctuations like the short range and medium range fluctuations as well as flatness parameter are studied. The estimated quality factors reflect the efficiencies with which the hadrons can be rejected from the data. Since some of these parameters are independent, the cuts may be applied in tandem and we demonstrate that the proton rejection efficiency of about 90% can be achieved. Use of density fluctuations is particularly suited for wavefront sampling observations and it seems to be a good technique to improve the signal to noise ratio.