I don't recall the exact details, but here is what I do remember ...
The sample - think a piece maybe 1" wide and 6" long - is held at an angle just shy of vertical. A large Bunsen burner is then placed under it, with the flame in contact with the lower edge for a brief period. Then the results are observed.
NOW ... for the bad news ..... this simple test has been shown to be essentially worthless in evaluating the actual fire behavior of materials. Only in the crudest manner does it provide any real means of comparing the hazards posed by the materials.
For example, nearly any wood product will ignite, and the flame will march up the face of the material. Eventually all the wood will be consumed. Simple, clearly flammable foam insulation will pass, though, as the material melts and falls to the bench as soon as it gets warm; the remaining sample is too far from the flame to ignite.
I do not recall the particular manner in which test results are 'scored.'
