A hefty part of designing a relational database is dividing the data elements into related tables. Once you're ready to start working with the data, you rely on relationships between the tables to pull the data together in meaningful ways. For instance, order information is useless unless you know which customer placed a particular order. By now, you probably realize that you don't store customer and order information in the same table. Instead, you store order and customer data in two related tables and then use a relationship between the two tables to view each order and its corresponding customer information at the same time. If normalized tables are a relational database's foundation, then relationships are the cornerstone.
The article continues at http://builder.com.com/article.jhtml?id=u00320030430ssh01.htm&page=1&vf=tt