In contrast to other motion entities that are attached to one or two parts, a 3D contact entity is comprised of one or two contact containers. A contact container contains one or more motion parts. In the case where a contact entity is defined by two contact containers, 3D contact is enabled between every part in the first container, and every part in the second container. In the case where a contact entity is defined by one contact container, 3D contact is enabled between every part in the container and every other part in the container.
When you define a contact entity, you enter parts in 1 or 2 selection sets. These selection sets generate the containers. If you only add parts in the first selection set, then only one container is generated. If you enter parts in two selections sets, then two container groups are generated.
Consider the example of the two containers shown next:

If a 3D contact object is attached to Container1 and Container2, contacts are detected between the following pairs of parts:
Part A and Part D
Part A and Part E
Part B and Part D
Part B and Part E
Part C and Part D
Part C and Part E
If a 3D contact object is attached only to Container1, contacts are detected between the following pairs of parts:
Part A and Part B
Part A and Part C
Part B and Part C
Consider another example of some balls bouncing in a vase:

This problem can be modeled using two 3D contact entities. One is attached to two contact containers. The first contact container contains the four balls, and the second contact container contains the vase. This enables contact between each of the four balls and the vase. The second 3D contact object is attached to a single container containing only the four balls. This enables contact between each ball and all the other balls.