.A current priority in Europe and the USA is the development of improved side collision dummies. This report presents the results of sled tests with three test subject types: cadavers (PMHS), EUROSID 1 and BIOSID. Twenty one (21) cadaver tests were performed and 9 dummy tests a piece. The left side of the test subjects were impacted under one of two different test conditions: 24 km/h rigid wall and 32 km/h padded wall. The cadavers were instrumented with a 12 thoracic, and triaxial pelvic accelerometer arrays. Thoracic deformation was calculated from rib accelerations. The dummies were instrumented in their standard formats, which included the ability to measure coronal plane thoracic deformation.For all test subject types and measurement locations the 3ms. acceleration standard deviations were low. Mean 3ms. accelerations showed no consistent relationship in magnitude between subject types. The measured dummy rib deformations were compared to the calculated cadaver deformations. The mean cadaver thoracic deformations for both test conditions were greater than those measured for the dummies. Comparing the dummies, at 24 km/h BIOSID demonstrated greater deformation and at 32 km/h EUROSID 1 exhibited the greater deformation. Standard deviations for dummy rib deformations were low, as expected cadaver values were high. The normalized shoulder panel forces for the dummy tests were compared to the ISO corridors, which in turn had been previously calculated from this cadaver data. EUROSID 1 data was in good agreement with the corridors for both tests, however BIOSID forces during 24 km/h impacts were too large. The dummies' general movement and impact posture influenced the resulting accelerations and forces. BIOSID developed a lean towards the impact panel during the pre-impact phase, which resulted in higher shoulder forces and lower pelvic forces. TTI and VC were calculated from the cadaver data. The best predictor of trunk AIS injury severity was VC calculated from acceleration at the 4th. rib combined with an age factor.