Major confusion exists with regard to the definition of patients with Raynaud's phenomenon; defining the patient and the phenomenon are reasonably straightforward, but variations in the definition of its primary and secondary forms have created a situation in which the same patient might be classified as primary by one group and secondary by another. The present essay is a proposal for the strict definition of Primary Raynaud's Phenomenon (PRP) formulated as a hypothesis amenable to experimental testing. This hypothesis is tested retrospectively on a group of 240 patients with Raynaud's phenomenon. The proposed criteria permit classification in 215 of 240 cases or 89%, leaving 25 patients difficult to classify at initial evaluation. Further testing of the hypothesis is encouraged.