Conconi et al. (1982) reported the development of noninvasive field test for anaerobic threshold (AT) based upon an observed deviation from the linear heart rate (HR)--running velocity (RV) relationship at high RV (HRdev). While the validity of the Conconi test has been debated (Heck and Hollmann 1992; Tokmakidis and Leger 1992), the reliability of the Conconi test has never been independently assessed in athletes performing the protocol outlined by Conconi. This study evaluated the reliability of the Conconi test in 15 well-trained male distance runners (22.5 +/- 3.3 yrs, 67.7 +/- 4.8 kg, VO2 peak 66.4 +/- 4.8 ml.kg-1.min-1) who performed a treadmill simulation of the Conconi test protocol twice within a 4-8 day period. The results were as follows: 6 subjects demonstrated HRdev in both Conconi tests, 5 subjects demonstrated HRdev in only one test, and in 4 subjects HRdev could not be discerned in either test. It was concluded that failure to determine a reproducible HRdev by subjective assessment in 9 of 15 subjects makes the Conconi test unsuitable for reliable evaluation of AT.