What is the difference between distributed and replicated database?

Distributed and replicated databases have a topology of having roles : 'masters' and 'slaves' even though in a replicated database schema, the distinction becomes a little blurred.

Distributed databases are normally those where information is updated on a master. The changes in the master are distributed out to a number of slave databases. The slave databases can be updated themselves but if they are, their information isn't distributed to any other database. Should the master have updated information that would contend with information in a slave, the master would take presidence.

Replicated databases are normally those in which there is the capability to post updates to both 'master' through to 'slave' as well as 'slave' through to 'master'.