The subject of this paper forms part of a broader effort to model the mechanics of grasping and fine-manipulation for robots. Grasping is the act of acquiring and holding (gripping) an object. Fine-manipulation is an extension of grasping to include control of the object using an end-effector such as a gripper or a hand. A mechanical model of grasping and manipulation forms the basis for controlling grippers and paves the way for robots that can make independent judgments about how to pick up and handle the objects they encounter. In this paper a procedure is developed for computing physical properties with which a grasp may be described. Among these properties are grip strength, stability, compliance and mobility. The results depend strongly on the interaction between the gripping surfaces and the object. For example, a grasp may be unstable when the fingertips are pointed, but stable for rounded fingertips. The analysis suggests that particular kinds of sensory information are especially useful in controlling a grasp and supports the notion that general grasping rules of thumb can be identified for use by robots. Originator- supplied key words include: Mechanical properties, Three dimensional, Fingers, Kinematics.