The toolkit implements each object cache as a Hashtable that maps the name of the cacheable object to a Vector that contains instances of the cacheable object. This implementation means that the size of an object cache can increase both in the number of types stored in it and in the number of instances of the types.