The rate of heat transfer to a specimen in The Fire Propagation Test has been measured. It rises as the test proceeds through a range corresponding roughly to that in the Spread of Flame Test (B.S. 476) and after about five minutes this rate of heat transfer becomes greater than that at the position of maximum spread for Class I material in the Spread of Flame Test. This presumably is one reason why The Fire Propagation Test distinguishes between low spread materials better than does the classification in the Spread of Flame Test. The rates of heat transfer in both tests are much higher than that necessary for the pilot ignition of untreated cellulosic materials but they may not be sufficient for the pilot ignition of some treated ones. They are, also, less than the heat transfer rate recorded over an experimental fire in a model corridor. The calculated initial rate of spread of fire along a wooden floor in this model when lined with a ceiling of cellulosic building board is well correlated with the fire propagation index of the ceiling board. Further work is in progress to assess whether the test is sufficiently sensitive for a wider range of materials.