A second-generation mechanical head-neck model was constructed, instrumented and subjected to pendulum impact tests against both the head and torso and directed from the front, rear and side. The response history of the system was measured by thirty channels of instrumentation including disk pressure transducers and muscle displacement gages in the neck, and a central accelerometer, intracranial pressure transducers and skull strain gages for the cranium and its contents. The kinematics of the unit was observed by an intermediate speed framing camera and the input was determined by a calibrated force transducer located at the contact point. It was found that peak head linear acceleration and velocity occur either during or immediately after the impact, with corresponding peak rotational values manifested somewhat later, but well before maximum head displacement. Head accelerations were similar, albeit slightly lower than in corresponding cases for an earlier model and displacement values were also similar until large extensions were reached. For rear head or frontal base impact, the head experienced a significant period of translation without rotation immediately after loading, and the system appears to respond more violently to side than to corresponding front or rear impact. The muscle beahvior, which support the findings from the head kinematics, is analyzed in detail and shows its strong influence on limiting head excursions, with strain values up to 40 percent. Disk pressure histories were similar to those found in tests on an earlier model with the highest values between T2 and C4, while the intercranial pressure exhibited more realistic values, about an order of larger magnitude.