wanting advice on finding out the root element in a w3c schema

i need a advice....................i am parsing a xml schema using java and trying to find out the root element in xml schema.................it is working quite well for almost 95% for all schemas............the problem is for some schemas some global elements are not referred in the schema , they r just defined and not referred anywhere..........so when parsing the global elements which r not referred anywhere are coming as root elements...........as far as i know any schema would have only one root element, but it is not the case here........i am getting multiple root elements............should i display tree structure for each root element i am getting........i have referred schema part 0,1,2.........i have not found ........it has mentioned only about 2 rules defining global elements such as (1) the element refer should not be placed int the global element and (2) min occurs and max occors(ranges) should not be placed in defining the global element...........so please tell me is there any logic to find out the root element other than the logic i am using.............i am not able cross this limitation as this is the bootleneck , b'cos the schema should reflect the xml document, and validating the xml document should be easier..............i think who is reading this got the point...............if not mail me back without hesitation...............