Castor-integration.md

@@ -145,4 +145,8 @@ When PEP's import routine encounters Castor data, it simply attempts to store it
 3. If the column names listed in step 2 are not to your liking, use [the `pepcli castor column-name-mapping` command](Using-pepcli#castor-column-name-mapping) to define [column name mappings](column-name-mappings). Repeat steps 2 and 3 until all column names are acceptable.
 4. Use [the `pepcli castor create-import-columns` command](Using-pepcli#castor-create-import-columns) to add import columns that do not yet exist. The command automatically places newly created columns into the `Castor` column group, making them writable to the import process. (Alternatively, use the `pepcli ama column` command to create and group columns manually.)
 
-When the configuration of the Castor study changes, different and/or additional columns may be needed when these data are imported into PEP. This requires Data Administrator to reconfigure import columns, i.e. perform the described steps again.
\ No newline at end of file
+# Limitations
+
+When the configuration of the Castor study changes, different and/or additional columns may be needed when these data are imported into PEP. This requires Data Administrator to reconfigure import columns, i.e. perform the [column configuration steps](#column-configuration) again.
+
+Note that PEP's import routine does not deal well with removals in the Castor study's configuration. For example, when a survey is removed from a survey package, the corresponding column names can no longer be listed or created by [the `pepcli castor` command](Using-pepcli#castor). But if data has been entered on the basis of the old configuration, the import process will still attempt to store it under the unlisted name. Castor and PEP administrators should take this into account when reconfiguring the Castor study and/or PEP's import columns.
\ No newline at end of file