One-hundred-and-one patients referred because of Raynaud's phenomenon (RP) were prospectively followed for a mean period of 42 months. At presentation they were screened for signs and symptoms of connective tissue disease (CTD) according to a detailed protocol. At presentation 37 patients had primary RP (PRP), nine had RP in combination with vascular occlusive disease (RP-VOD), 25 had one symptom of a CTD (questionable PRP), 13 had two or more symptoms (undifferentiated CTD, UCTD) and 17 had definite CTD. Progression from one of these groups to another was seen in 24 patients and from PRP, RP-VOD or questionable PRP towards a (U)CTD was seen in 19 patients. Patients with one sign of CTD showed a high tendency (56%) to develop CTD. The presence of ANA as detected by immunofluorescence and by immunoblotting at the start of the study was associated with the future development of symptoms of CTD; positive predictive value 65% and 71% and negative predictive value 93% and 83%, respectively. ANA-testing by immunoblotting was of special help in predicting the development of scleroderma, the CREST syndrome and mixed connective tissue disease. In conclusion, testing for ANA by indirect immunofluorescence helps to discriminate between patients with persisting PRP and those who will develop a CTD, while testing for ANA by the immunoblotting technique helps to predict the development of a specific CTD.