An intercomparison on thoron ((220)Rn) measurement was carried out between National Institute of Radiological Sciences, Japan (NIRS), and New York University School of Medicine, USA (NYU). The measurements of (220)Rn concentration at NIRS and NYU were performed by using the scintillation cell method and the two-filter method, respectively, as the standard measurement method. Three types of alpha track detectors based on passive radon ((222)Rn)-(220)Rn discriminative measurement technique were used: Raduet and Radopot detectors were used at NIRS, and four-leaf detectors were used at NYU. In this study, the authors evaluated (220)Rn concentration variation in terms of run for exposure, measurement method, and exposure chamber. The detectors were exposed to (220)Rn gas with approximately 15 kBq m(-3) during the period from 0.75 to 3 d. As a result, the variation of each measurement method among these exposure runs was comparable to or less than that for the two-filter method. Agreement between the standard measurement methods of NIRS and NYU was observed to be about 10%, as is the case with the passive detectors. The Raduet detector showed a large variation in the detection response between the NIRS and NYU chambers, which could be related to different traceability.