Sometimes a tester might be assigned to reproduce some issues from older test runs to see whether the issue still exist, and by doing so, clean up the project issues backlog.
The reason for this is that our clients only fix issues that are important to them not all the issues reported in a project. During development work, whole parts of applications, that affect issues which have previously been reported, can be scrapped or impacted without being "fixed".
The steps of reproducing are the same as when doing (fixed) issue verification with this exception: close as Cannot reproduce if the issue is not reproducible.
Think you've got the hang of it? Don't forget to read about when and how to reproduce issues.