Abstract The theory that β‐ketonitrile structures initiate discoloration reactions in polyacrylonitrile has been tested by a number of methods. Using radiochemical techniques the numbers of carbonyl groups in various polymers have been measured as have the amounts of combined water. A model β‐keto compound, 1,3,7,‐tricyanoheptan‐4‐one was synthesized and its reactions studied. None of these experiments supported the idea that β‐ketonitrile structures are of primary importance in initiating the degradation of polyacrylonitrile.