why the long read operations in secondary replica need the ghost records from the primary replica ?

Based on the article below ,it seems the cleanup of ghost records on the primary replica can be blocked by transactions secondary replicas, as the secondary replica read operations need these ghost records from the primary replica , so my question is why need them?

"Because read operations are mapped to snapshot isolation transaction level, the cleanup of ghost records on the primary replica can be blocked by transactions on one or more secondary replicas. The ghost record cleanup task will automatically clean up the ghost records for disk-based tables on the primary replica when they are no longer needed by any secondary replica.  "

http://technet.microsoft.com/en-us/library/ff878253.aspx