The objective of this study was to determine the strength of bonded amalgam restorations involving an entire cusp. Four groups were established, each containing 13 caries-free, recently extracted molar teeth of similar size and anatomic form. The teeth were stored in 70% alcohol prior to the removal of a cusp and the cutting of a conventional preparation with an occlusal lock. Group A were restored with amalgam only, Group B with amalgam and a pin, Group C with amalgam and Amalgambond (4-META) and Group D with amalgam, a pin and Amalgambond. The teeth were blocked in stone and stored for 7 days in water at 37 degrees C. Using an Instron, a force was delivered at right angles to the inclined cuspal plane of the amalgam at its midcuspal point at a crosshead speed of 5 mm per minute. The load at failure was recorded and the assemblies recovered for fracture analysis using SEM. Group A failed at 33.61 +/- 15.17 Kg, Group B at 32.28 +/- 12.10 Kg, Group C at 48.53 +/- 21.93 Kg and Group D at 47.84 +/- 14.79 Kg. No statistically significant difference existed between Groups A and B, and Groups C and D. A significant difference (P < 0.002) existed between Groups A and C and B and D. Failure commonly occurred between cusp and occlusal lock or within the cusp when a pin was used. Bonded restorations with a pin incurred cohesive root failure. It was concluded that the strength of the amalgam restorations benefitted significantly from bonding but the Amalgambond system may be contraindicated for use with pins due to root fracture.