SQL ROW_NUMBER Behavious

I'm seeing some strange behaviour with the Row_Number function in SQL Server 2008. Given the following code:

SELECT  ROW_NUMBER() OVER(PARTITION BY ACPL.Identifier ORDER BY ACPL.ClassName,
 ACPL.Identifier), 
	ROW_NUMBER() OVER(ORDER BY ACPL.ClassName, ACPL.Identifier),
		ACPL.Identifier, ACPL.ClassName, Description, 
        MAX(UCAS.Initials), ' '
	FROM MarkingRecords..vw_KRC_SummerProfileSubmittedACPLNew ACPL
	JOIN MarkingRecords..vw_KRC_SummerProfileSubmittedUCAS UCAS
	ON ACPL.ClassName = UCAS.ClassName AND
	ACPL.Identifier = UCAS.Identifier
	GROUP BY ACPL.Identifier, ACPL.ClassName, Description

Most of the numbers in the first column are correct, but at some points the row number does not return to 1 when the partition is applied. For example:

1      812      4278      1:BLA:D1            NFG
2      813      4278      1:BLA:D1            NFG
3      814      4278      1:BLA:D1            NFG
4      815      4278      1:BLA:D1            NFG
5      816      4278      1:BLA:D1            NFG
4      817      4314      1:BLA:D1            NFG
5      818      4314      1:BLA:D1            NFG
6      819      4314      1:BLA:D1            NFG

As you can seem the third column (Identifier) changes from 4278 to 4314, but the Row_Number starts from 4. Any ideas why this would happen? I've used Row_Number on many occasions before without any problems, although this is the first time that I've used two in one query (Bad database design; there is no unique identifier column in the database tables, also why there are reserved keywords as column names!)