Results of investigations of the factor VIII (FVIII) of a patient with an unusual variant form of von Willebrand's disease (vWD) are presented. A two-peak crossed-immunoelectrophoresis (CIE) pattern was seen when fresh plasma was electrophoresed, but the CIE pattern became normal by incubating the plasma at 37 degree C for more than 72 hr. The two peaks on CIE were separated by cryoprecipitation: the slow-moving peak precipitating and the fast-moving forms of FVIII remaining in the cryosupernate. An additional protein band was seen on multimeric analysis of FVIII. The platelet-rich plasma (PRP) from this patient did not respond to ristocetin, but agglutinated normally in response to botrocetin. Multimeric and CIE analysis of the FVIII post agglutination and 125I-FVIII binding studies to normal formalin-fixed platelets indicated that this patient's FVIII interacted normally with botrocetin but failed to interact with ristocetin. These data strongly suggest that the sites on the FVIII molecule or the multimeric forms involved for ristocetin and botrocetin are different and that the ristocetin reaction is more closely aligned to the physiologic function of FVIII.