Experiments were conducted using a laboratory filter column to establish the best wash regime and establish the nature, and extent, of filter medium fouling for filter sand taken from a water treatment plant in London, United Kingdom. A range of different air-water combinations were investigated starting from those currently used and spanning a range of collapse-pulsing combinations. Chlorinated backwash water was also used to assess whether this could be effective in removing the fouling layer. The maximum existing air and water flow rates used at the works were found to be insufficient to give collapse-pulsing and did not show strong bed agitation or good cleaning in the laboratory. SEM analysis showed a biofilm on the sand grains, with bacteria and EPS matrix growing over grain surfaces and a lot of bacteria in grain crevices. This biofilm was not being removed by the existing backwash regime. Best cleaning in the laboratory was by collapse-pulsing, with a water velocity at 60% of vmf and an air velocity of 48m/h. This resulted in the cleanest grain surfaces, particularly after 20ppm chlorine was employed in combination with the air-water wash. The filter start-up head loss gradient was reduced by 18%. The sand discoloration did not change significantly during the experiments, even by employing 50ppm chlorine in the wash water. SEM analysis demonstrated that the dark brown colour of sand was due to the biofilm, dirt and some small amounts of manganese. This established biofilm was very strongly attached to the sand grains. Includes 4 references, figures.